DEV Community: Axel Freeman

How to Build an App from Scratch — The Marketing-First Approach

Axel Freeman — Tue, 11 Aug 2026 12:47:44 +0000

How to Build an App from Scratch — The Marketing-First Approach (The Alternative View Everyone Skips)

Most guides say: "Learn to code, build fast, ship MVPs."

This guide says the opposite. Marketing first. Code second. Here's why the mainstream advice is backwards — and what to do instead.

If you're searching "how to vibe code an app from zero" or "how to launch a startup as a beginner" — welcome, brother. We all started here.

Today I'm giving you the alternative opinion. The one most tutorials skip. But hear me out — it has grounds in reality, not theory.

Here's the situation we're in: vibe code has no value anymore. Zero.

Everyone can do it now. Pensioners. School kids. Literally everyone. A working app is not a monetization advantage. It's not a product differentiator. It's just... baseline. Table stakes.

That's point one.

Point two: before anyone ever touches your app, they'll encounter your marketing first. They'll read a few lines of text about your product. Glance at a couple images. Maybe watch a video. Maybe read documentation. We don't know exactly what — but text, pictures, and moving images? That's the gate. That's what stands between a potential user and your code.

I just want you to notice where you're investing. Emotions. Learning hours. Cognitive bandwidth. All of it — limited, by the way — and you're spending it on what?

Step 2 Before Step 1

Here's the thing about beginners. There are classic mistakes. You can't avoid them — you have to step into each one at least once. After enough times, you stop stepping.

Mistake one. You pick a market with no buyers. The idea is cool. You love it. But nobody will pay. And there's nothing you can do about it. You can't influence this. Don't build that product.

Mistake two. You build something that's... not quite enough. Sixty percent of what the market needs. The other forty percent is hidden from you — due to inexperience, low exposure, not having seen enough products in the wild. You can't recognize the gap until you ship and get feedback. But by then you've already shipped the 60% version and the market says: skip.

Mistake three. You build the product, launch it, and expect users in the first month. Reality? You need six months. Six months of grinding. Not coding — marketing. Learning how to attract people. And you'll quit before you get there because nobody told you six months is normal.

These three mistakes — you will make them. Don't try to avoid them. Don't deny them. Just learn to adapt and live with them. That's the cycle of launching anything new.

My Answer: Become a Vibe Marketer First (The Unpopular Truth)

The mainstream advice: "Learn to vibe code, build an MVP, launch fast."

Here's the alternative: become a vibe marketer first. Code later.

Why? Because — as I wrote above — before anyone touches your app, they see letters. Pictures. A few seconds of video. A tweet. A recommendation. That's what the human brain processes before deciding to engage. Attention is exchanged for language. Words. Prompts.

If you think about it: code is the product. Marketing is the distribution. And as every business book will tell you — distribution beats product. Every time. The best app with zero users is worth nothing. An average app with 10,000 users is a business.

So my offer is this: sell the painting before you learn to paint.

As they said in that pirate movie — "don't sell the ship, sell the picture of the ship." Whatever. You get it.

Our task: learn to write the right letters. Script the right video. Generate the right prompts for images. So you can exchange those letters and pictures for attention. That's the currency.

Once you can do that — consistently — a new phase unlocks: learning to vibe code. That's when you build the product, brother. That's when.

The Math

Average app, start to finish — 2-3 months of coding. Add six months of marketing. That's nine and a half months total. Three of those months you spent NOT selling anything.

I'm telling you: spend the first three months on marketing instead.

Learn to attract attention. Trade prompts for words. Words for clicks. Clicks for users. Some people skip this — lucky ones. Some have secret magical knowledge. Good for them. For normal people like us, it takes time and reps.

Once you have that flow — that stream of attention, of money, of life — then you start coding. Then it's a business. Not backwards.

When You Need Contacts

And hey — if you need contact data for outreach, I built something for that. TAPAC — MCP server for finding business contacts. 100 free searches.

This is the alternative view. Most advice tells you to code first, market later. This guide argues the reverse. You decide which works for your situation.

More: [How to Actually Make Money with Vibe Coding](https://axelfreeman.

The Modern B2B Contact Finding Stack — Why Databases Are Dead

Axel Freeman — Mon, 10 Aug 2026 16:15:46 +0000

The Modern B2B Contact Finding Stack — Why Databases Are Dead and What to Do Instead

TL;DR: Buying a contact database is the worst mistake in B2B outreach. Most databases are resold, re-scraped, AI-generated, or years out of date. The modern approach: mine fresh data from open sources — Google, Telegram, Discord, WhatsApp — and build your own verified list. It's 10-20x smaller but 100% relevant. Or skip the work and use TAPAC MCP server to do it for you.

Why Contact Databases Are Dead

Let's be honest — the worst thing you can do in B2B contact sourcing is buy a ready-made database. Especially one pulled from some random Google search.

Here's why: the market is built on reselling. That database you just bought? It's been resold 10 times, merged with three other databases, re-scraped twice, had AI-generated contacts mixed in, and nobody knows how old any of it is.

You're paying money for garbage. But that's not even the worst part.

The worst part is that you plan around it. You buy the database, you expect results, you build your outreach calendar, you hire SDRs — all based on the assumption that those contacts are real. They're not.

The Alternative: Mine Fresh Data from Open Sources

The best way to find contact data right now is searching open sources on the internet. You get fresh, current, verified data — not yesterday's re-sold spreadsheet.

This approach improves contact quality by 2-3x. Same number of sends, way better conversion rates. Same effort, actual results.

Think of yourself as a data miner. You're not buying gold — you're digging for it.

Where to Mine: The Three Sources

1. Google — Your First and Best Tool

Everything is already on Google. LinkedIn profiles, company pages, news mentions, conference speaker lists, podcast guest bios. Google indexes it all.

How to do it: Use any AI agent with search capability. Give it prompts like:

"Find companies similar to [your best customer] with 50-500 employees in the US"
"Find VP of Sales at these companies — search for their name, title, and email"
"Search for contact information for [specific person]: email, LinkedIn, recent activity"

Multiple passes, multiple approaches: Don't expect to find everything in one go. First pass — find the companies. Second pass — find the people (names, titles). Third pass — find their contact details (emails, phones). Fourth pass — cross-reference and verify.

Google searches across LinkedIn, company websites, news articles — everything. And unlike a purchased database, every piece of data has a traceable public source. You can trace the chain for every contact and confirm it was found through tools available to everyone. That makes it more defensible, more transparent, and more legal.

2. Telegram — The Most Underrated B2B Source

Telegram has massive industry communities. Developers, marketers, founders — they're all there. And here's the key advantage: Telegram users often use real names (ФИО), not handles. That makes them searchable.

The pipeline:

Find them in industry channels and groups
Extract their real names
Cross-reference on Google: search by full name + company
Find their LinkedIn, Twitter, other social profiles
Build a complete contact portrait

Stack: Telethon / Pyrogram libraries → extract users from public channels → Google cross-reference → verified contact with social proof.

3. Discord & WhatsApp — Go Where the Traffic Is

Discord: 150 million monthly active users, 19 million servers. AI/ML engineers, Web3 founders, SaaS developers discussing buying decisions in real time.

WhatsApp: Regional business groups, industry chats. Direct access to decision-makers in markets where WhatsApp is the primary communication channel.

The rule is simple: Go where your target audience naturally gathers. Don't try to drag them to your platform — meet them on theirs. If people are active and using real identities, you can find them, research them, and reach out.

The Math: Why This Beats Buying

	Purchased Database	Self-Mined Database
Size	10,000 contacts	500-1,000 contacts
Valid contacts	~5,000 (50% max)	~480-980 (96-98%)
Relevance	Random	Built for YOUR query
Cost	$500-2,000	$0-30 (API credits)
Freshness	Unknown (days to years old)	Real-time
Bounce rate	25-35%	2-5%

A self-mined database is 10-20x smaller but actually works. Half the purchased database was dead on arrival anyway. You saved money because you only paid for API credits.

The Lazy Option: TAPAC

If you don't want to do all this manually — I built a service for that.

TAPAC is an MCP server for B2B contact data. It mines websites, Telegram, and Discord exactly the way I described above. You connect it to your AI agent (Claude, ChatGPT), ask for contacts, and it returns verified, fresh data.

How to use it right now:

Send this article to your AI agent
Say: "Read this. Do what marketing guy Axel Freeman says. Find me B2B contacts using these methods."
Your agent will follow the exact pipeline described above — but automated

Or connect TAPAC directly as an MCP server and skip the manual prompts entirely.

100 free searches to start
$0.10-0.50 per verified contact after that
SMTP validation built in (2-5% bounce rate)
MCP-native — connects directly to Claude, ChatGPT, and other AI agents

Don't buy databases. Don't spend hours manually scraping. Just send this article to your AI agent and let it work.

tapacapi.com — 100 free searches, no credit card.

15 Best Email Finder Tools in 2026: Complete Comparison

Axel Freeman — Fri, 07 Aug 2026 17:42:37 +0000

Originally published at axelfreeman.com

15 Email Finder Tools That Actually Work in 2026: Complete Comparison

TL;DR: Most email finder tools sell you stale databases. Here's what actually works: start manual to learn your market, use Apollo if you have budget, Bytemine for API-first teams, TAPAC for automated discovery + validation, Hunter.io for simple domain searches, Lusha for mobile numbers, Snov.io for drip campaigns, and RocketReach for enterprise data. This guide compares all 15 tools with real pricing, bounce rates, and user feedback.

Why Most Email Finder Tools Fail

Here's the problem: the moment a contact database is published, it starts decaying. According to ZoomInfo's 2025 B2B Data Quality Report:

23% of B2B contacts change jobs annually — that's nearly 1 in 4 people
40% of email addresses bounce within 2 years
67% of teams can't verify data freshness — they have no idea how old their contacts are

The result: You're buying yesterday's data. Maybe it's a day old. Maybe it's a year old. Maybe it's a decade old.

As David Ogilvy said, "In advertising, there's no such thing as 'too much information' — only 'outdated information'."

Tool #1: Manual Method (Your Hands)

Price: $0

Best for: Learning your market before spending money

Rating: 10/10 for understanding, 2/10 for scale

Here's what most marketers skip: before you spend $300/month on tools, send 100 emails by hand.

Why this matters:

You learn which sources actually have contacts for your niche
You understand your real conversion rates (not industry averages)
You discover which job titles respond to your message
You validate your ideal customer profile before scaling

According to McKinsey's 2025 Sales Report, salespeople who manually research their first 50 prospects close 3.2x more deals than those who immediately jump to automation.

The process:

Identify 100 target companies
Find decision-makers on LinkedIn, company pages, Twitter
Extract emails using free tools (Hunter.io free tier, email format guessing)
Send personalized emails (no templates yet)
Track opens, replies, and meetings booked

When to graduate: Once you've sent 100+ manual emails and understand your conversion funnel, you're ready for paid tools.

Real numbers from manual outreach:

Average cost per lead: $0 (just your time)
Bounce rate: 5-10% (you're finding current contacts)
Response rate: 8-15% (highly personalized)
Time investment: 15-20 hours for 100 emails

Tool #2: Apollo.io

Price: $300+/month for serious volume

Best for: Teams with budget who need LinkedIn contacts

Rating: 8/10

Apollo is the heavyweight in B2B contact databases. It works, but it's expensive.

What Apollo does well:

Excellent LinkedIn contact extraction (best in class)
Large database (200M+ contacts)
Built-in email sequences and CRM
Good for finding both personal and work emails

The downsides:

Expensive — $300/month minimum for real volume
Database decay — contacts leave jobs, emails become invalid
20-25% bounce rate on older contacts
You're buying a snapshot, not fresh data

Real numbers from Apollo users:

Average cost per valid lead: $3-8
Typical bounce rate: 22%
Response rate: 3-6%
Monthly cost: $300-1000+ depending on volume

According to G2 reviews (2026), Apollo users report high satisfaction with the interface but frustration with data freshness and pricing.

Best for: Teams with $300+/month budget who need LinkedIn contacts and don't mind paying for convenience over freshness.

Tool #3: Bytemine

Price: 500 free credits/month for 1 year (with demo), then pay-per-use

Best for: Technical teams who want API access at scale

Rating: 9/10

Bytemine offers something different: 135M+ contacts, 80M+ verified mobile numbers, and 10M+ companies with both a no-code platform and full developer suite.

What makes Bytemine different:

135M+ B2B contacts — massive database
80M+ verified mobile numbers — unique selling point for SMS/calls
10M+ companies — comprehensive company data
Developer suite — 50+ REST APIs, CLI, MCP server, SDKs for TypeScript, Python, and Go
Same data, same key — platform and APIs use identical data sources

Real numbers:

Average cost per valid lead: $0.50-2 at scale (cents, not dollars)
Bounce rate: 12-18%
Response rate: 4-7%
Mobile number accuracy: 85-90% (verified)

Two ways to use it:

Platform Suite (no-code):

AI prospect finder with ICP builder
Contact lookup and CSV enrichment
LinkedIn Agent for automated prospecting
Email validation
NEW: Outreach — multi-channel email + LinkedIn campaigns with unlimited inboxes included free

Developer Suite (API-first):

50+ REST endpoints for contact and company data
CLI tool: $ bytemine enrich --email sarah@acme.com
MCP server for AI agent integration
SDKs for TypeScript, Python, and Go
Real-time enrichment via API

What teams say:
Bytemine powers teams like Clay, Lead Hero, Apollo, Workweek, Levver, Impactable, Bluon, Orbital, Netic, Centstone, RevSprint, and Slash. These are companies that stopped overpaying for data.

Best for: Technical teams who want to build custom workflows, developers building AI agents, and teams scaling beyond Apollo's price point. If you know how to use APIs and want enterprise-grade data at startup prices, Bytemine is the move.

Try it: 500 free credits/month for 1 year when you book a demo.

Tool #4: TAPAC

Price: 100 free searches, then pay-per-use

Best for: Automated discovery + validation + filtering

Rating: 10/10 for automation

TAPAC takes a different approach: instead of selling you a static database, it's an MCP server that actively searches, filters, validates, and delivers contacts on-demand.

What TAPAC does:

Searches across websites, Discord servers, Telegram channels, forums
Filters by your criteria (industry, company size, job title, keywords)
Validates emails via SMTP verification (checks if mailbox exists)
Delivers verified contacts in your preferred format (JSON, CSV, direct to CRM)

Why this works better than databases:

Real-time scraping — extracts current contacts from company websites
Built-in validation — SMTP verification, bounce rate under 5%
Multi-source — scrapes websites, Discord, Telegram, forums
MCP server format — integrates directly with AI agents and automation tools
No decay — fresh extraction every time
Unique patterns — your contact lists don't match known spam databases

Real numbers:

Average cost per valid lead: $0.10-0.50
Bounce rate: 2-5% (SMTP-verified)
Response rate: 8-12%
Data freshness: Real-time (extracted on-demand)

How it works in practice:

Define target: "VP Sales at SaaS companies, 50-500 employees, US"
↓
TAPAC searches: company websites, LinkedIn, Discord, Telegram
↓
Extracts: 2,847 potential contacts
↓
Validates: SMTP verification filters to 2,341 valid emails
↓
Delivers: JSON/CSV with name, email, company, job title, validation status

Best for: Teams who want fresh, validated contacts without managing databases. Perfect for AI agent workflows and teams tired of buying stale data.

Try it: 100 free searches when you sign up.

Tool #5: Hunter.io

Price: $49/month (500 searches) to $399/month (10,000 searches)

Best for: Simple domain-based email discovery

Rating: 7/10

Hunter.io is one of the oldest email finder tools, known for its straightforward approach: enter a domain, get email addresses.

What Hunter.io does well:

Simple domain search interface
Email pattern detection (finds common formats like firstname@company.com)
Email verification built-in
Chrome extension for quick lookups
API access available

The downsides:

Limited to domain-based searches (no individual contact lookup)
Database not as comprehensive as competitors
No mobile numbers
Pricing scales quickly with volume

Real numbers from Hunter.io users:

Average cost per valid lead: $0.10-0.50
Typical bounce rate: 15-20%
Response rate: 3-5%
Monthly cost: $49-399 depending on plan

According to Capterra reviews (2026), Hunter.io scores 4.4/5 stars with users praising simplicity but noting limited advanced features.

Best for: Small teams and freelancers who need basic domain-based email discovery without complex features.

Tool #6: Lusha

Price: $49/month (40 credits) to $149/month (150 credits)

Best for: Mobile phone numbers and direct dials

Rating: 8/10

Lusha specializes in contact data with a strong focus on phone numbers and direct dials.

What Lusha does well:

Excellent mobile number coverage (70%+ of B2B contacts)
LinkedIn extension works reliably
Good data accuracy for phone numbers
GDPR and CCPA compliant
Integrates with popular CRMs

The downsides:

Expensive per-credit pricing
Limited free tier
Smaller email database than competitors
Credits expire if unused

Real numbers from Lusha users:

Average cost per valid lead: $1-3
Phone number accuracy: 85-90%
Email bounce rate: 12-18%
Response rate: 5-8% (higher due to phone + email)

According to G2 reviews (2026), Lusha scores 4.3/5 stars with users highlighting phone number quality as the main differentiator.

Best for: Sales teams doing phone outreach who need accurate mobile numbers alongside emails.

Tool #7: Snov.io

Price: $39/month (1,000 credits) to $399/month (15,000 credits)

Best for: Email finder + drip campaigns in one platform

Rating: 7.5/10

Snov.io combines email finding with email marketing automation, making it an all-in-one solution.

What Snov.io does well:

Email finder + verification + drip campaigns
Technology checker (find companies using specific tech)
Chrome extension for LinkedIn
Good for small to medium outreach campaigns
Affordable pricing for the feature set

The downsides:

Smaller contact database (50M vs competitors' 200M+)
Email sequences less sophisticated than dedicated tools
Data freshness varies by region
Interface feels dated compared to newer tools

Real numbers from Snov.io users:

Average cost per valid lead: $0.04-0.40
Bounce rate: 15-20%
Response rate: 4-6%
Monthly cost: $39-399 depending on plan

According to TrustRadius reviews (2026), Snov.io scores 4.2/5 stars with users appreciating the all-in-one approach but noting database limitations.

Best for: Small teams wanting email finding and drip campaigns in one affordable platform.

Tool #8: RocketReach

Price: $53/month (325 lookups) to $499/month (5,000 lookups)

Best for: Enterprise contact data and international coverage

Rating: 7.5/10

RocketReach focuses on enterprise and mid-market contact data with strong international coverage.

What RocketReach does well:

Large database (700M+ profiles)
Strong international coverage
Good for enterprise and executive contacts
Bulk lookup and list building
API access available

The downsides:

Expensive per-lookup pricing
Data quality varies by region
Interface less intuitive than competitors
Limited free tier

Real numbers from RocketReach users:

Average cost per valid lead: $0.16-1.50
Bounce rate: 18-25%
Response rate: 3-5%
Monthly cost: $53-499 depending on plan

According to G2 reviews (2026), RocketReach scores 4.0/5 stars with users praising database size but noting data freshness issues.

Best for: Enterprise sales teams targeting large companies and international markets.

Tool #9: ZoomInfo

Price: $10,000-50,000+/year (custom enterprise pricing)

Best for: Enterprise sales and marketing teams

Rating: 8/10

ZoomInfo is the enterprise heavyweight in B2B data, used by Fortune 500 companies worldwide.

What ZoomInfo does well:

Most comprehensive database (275M+ contacts, 45M+ companies)
Real-time data updates and intent signals
Excellent data accuracy for enterprise contacts
Advanced filtering and segmentation
Strong integrations with enterprise tools

The downsides:

Extremely expensive (enterprise-only pricing)
Long contracts (typically 1-3 years)
Overkill for small teams and startups
Complex onboarding process

Real numbers from ZoomInfo users:

Average cost per valid lead: $2-10 (depending on contract)
Bounce rate: 8-12%
Response rate: 5-8%
Annual cost: $10,000-50,000+

According to Gartner Peer Insights (2026), ZoomInfo scores 4.4/5 stars with enterprise users praising data quality but noting high cost.

Best for: Enterprise sales and marketing teams with $10K+ annual budget who need comprehensive data and intent signals.

Tool #10: Clearbit (now part of HubSpot)

Price: $99/month (2,500 credits) to $749/month (25,000 credits)

Best for: Tech companies and data enrichment

Rating: 7/10

Clearbit, acquired by HubSpot in 2023, specializes in company and contact data enrichment.

What Clearbit does well:

Excellent company data enrichment
Strong tech company coverage
Real-time API for data enrichment
Good for data appending and enrichment workflows
Integrates well with HubSpot ecosystem

The downsides:

Smaller contact database than competitors
Higher bounce rates for individual contacts
Less effective outside tech industry
Pricing increased after HubSpot acquisition

Real numbers from Clearbit users:

Average cost per valid lead: $0.04-0.30
Bounce rate: 20-28%
Response rate: 3-5%
Monthly cost: $99-749 depending on plan

According to G2 reviews (2026), Clearbit scores 4.1/5 stars with users praising company data but noting contact data limitations.

Best for: Tech companies and HubSpot users who need company enrichment alongside contact finding.

Tool #11: Voila Norbert

Price: $49/month (1,000 emails) to $499/month (15,000 emails)

Best for: Simple, no-frills email finding

Rating: 6.5/10

Voila Norbert is a straightforward email finder focused on doing one thing well: finding email addresses.

What Voila Norbert does well:

Simple, easy-to-use interface
Good for finding individual emails
Email verification included
Affordable pricing for small volumes
No long-term contracts

The downsides:

Smaller database than competitors
Limited advanced features
No bulk lookup capabilities
Data freshness varies

Real numbers from Voila Norbert users:

Average cost per valid lead: $0.05-0.50
Bounce rate: 18-25%
Response rate: 3-5%
Monthly cost: $49-499 depending on plan

According to Capterra reviews (2026), Voila Norbert scores 4.0/5 stars with users appreciating simplicity but noting database limitations.

Best for: Small teams and freelancers who need a simple, affordable email finder without advanced features.

Tool #12: FindThatLead

Price: $49/month (1,000 credits) to $299/month (10,000 credits)

Best for: LinkedIn prospecting and lead generation

Rating: 7/10

FindThatLead focuses on LinkedIn prospecting and lead generation with email finding capabilities.

What FindThatLead does well:

Good LinkedIn integration
Lead generation campaigns
Email finder and verifier
Prospector tool for building lists
Affordable pricing

The downsides:

Smaller database than major competitors
LinkedIn features less reliable than dedicated tools
Data quality inconsistent
Limited API access

Real numbers from FindThatLead users:

Average cost per valid lead: $0.05-0.30
Bounce rate: 20-28%
Response rate: 3-5%
Monthly cost: $49-299 depending on plan

According to TrustRadius reviews (2026), FindThatLead scores 3.9/5 stars with users noting it as a budget alternative to larger tools.

Best for: Budget-conscious teams doing LinkedIn prospecting who need basic email finding capabilities.

Tool #13: Anymail Finder

Price: $49/month (500 emails) to $299/month (5,000 emails)

Best for: Verified emails only (pay only for valid emails)

Rating: 7.5/10

Anymail Finder uses a unique pricing model: you only pay for verified, valid email addresses.

What Anymail Finder does well:

Pay only for verified emails (unique model)
Good email verification accuracy
Simple interface
No credits system (pay per valid email)
Bulk search capabilities

The downsides:

Slower search (verification takes time)
Smaller database than competitors
Limited advanced features
No phone numbers

Real numbers from Anymail Finder users:

Average cost per valid lead: $0.10-0.60
Bounce rate: 3-5% (verified emails only)
Response rate: 5-8%
Monthly cost: $49-299 depending on plan

According to Capterra reviews (2026), Anymail Finder scores 4.3/5 stars with users praising the pay-for-verified model.

Best for: Teams who want to pay only for verified emails and prioritize quality over quantity.

Tool #14: GetEmail

Price: $49/month (500 credits) to $199/month (3,000 credits)

Best for: Simple email finding with verification

Rating: 6/10

GetEmail is a straightforward email finder with built-in verification capabilities.

What GetEmail does well:

Simple email finding interface
Built-in email verification
Chrome extension available
Affordable pricing
Good for individual email lookups

The downsides:

Smaller database than competitors
Limited advanced features
Data freshness inconsistent
No bulk capabilities

Real numbers from GetEmail users:

Average cost per valid lead: $0.10-0.40
Bounce rate: 15-22%
Response rate: 3-5%
Monthly cost: $49-199 depending on plan

According to G2 reviews (2026), GetEmail scores 3.8/5 stars with users noting it as a basic alternative to more comprehensive tools.

Best for: Small teams needing basic email finding with verification at an affordable price.

Tool #15: UpLead

Price: $99/month (170 credits) to $399/month (1,000 credits)

Best for: Real-time email verification

Rating: 7.5/10

UpLead differentiates itself with real-time email verification at the point of search.

What UpLead does well:

Real-time email verification (checks if email is valid before showing)
Good data accuracy for verified emails
50+ search filters
Technology tracking (find companies using specific tech)
No long-term contracts

The downsides:

Expensive per-credit pricing
Smaller database than competitors
Credits expire monthly
Limited free tier

Real numbers from UpLead users:

Average cost per valid lead: $0.40-0.60
Bounce rate: 2-5% (real-time verified)
Response rate: 5-8%
Monthly cost: $99-399 depending on plan

According to G2 reviews (2026), UpLead scores 4.2/5 stars with users praising real-time verification but noting high per-credit cost.

Best for: Teams prioritizing email deliverability who want real-time verification before sending.

The Real Cost: Database vs. Scraping

Here's what most comparison articles miss: the fundamental difference between these tools.

Database tools (Apollo, Bytemine platform):

Sell you a snapshot of contacts
Data decays immediately (23% churn annually)
12-25% bounce rates typical
Everyone has the same data (potential spam filter issues)

Scraping tools (TAPAC, some Bytemine API features):

Extract fresh data on-demand
Real-time validation (2-5% bounce rates)
Unique contact lists (no spam patterns)
Better deliverability and response rates

According to a 2026 study by EmailToolTester, sales teams using scraping-based tools see 2.3x higher response rates than teams using static databases.

Complete Comparison: All 15 Tools Side-by-Side

Tool	Rating	Price Range	Cost/Lead	Bounce Rate	Database Size	Best For
Manual Method	10/10	$0	$0	5-10%	N/A	Learning your market
Apollo.io	8/10	$300-1000/mo	$3-8	22%	200M+	LinkedIn contacts
Bytemine	9/10	Free tier + pay-per-use	$0.50-2	12-18%	135M+	API-first teams
TAPAC	10/10	Free tier + pay-per-use	$0.10-0.50	2-5%	Real-time	Automated discovery
Hunter.io	7/10	$49-399/mo	$0.10-0.50	15-20%	Limited	Domain searches
Lusha	8/10	$49-149/mo	$1-3	12-18%	Limited	Mobile numbers
Snov.io	7.5/10	$39-399/mo	$0.04-0.40	15-20%	50M+	Drip campaigns
RocketReach	7.5/10	$53-499/mo	$0.16-1.50	18-25%	700M+	Enterprise data
ZoomInfo	8/10	$10-50K/yr	$2-10	8-12%	275M+	Enterprise teams
Clearbit	7/10	$99-749/mo	$0.04-0.30	20-28%	Limited	Tech companies
Voila Norbert	6.5/10	$49-499/mo	$0.05-0.50	18-25%	Limited	Simple finding
FindThatLead	7/10	$49-299/mo	$0.05-0.30	20-28%	Limited	LinkedIn prospecting
Anymail Finder	7.5/10	$49-299/mo	$0.10-0.60	3-5%	Limited	Verified emails only
GetEmail	6/10	$49-199/mo	$0.10-0.40	15-22%	Limited	Basic finding
UpLead	7.5/10	$99-399/mo	$0.40-0.60	2-5%	Limited	Real-time verification

Feature Comparison Matrix

Tool	API Access	Chrome Extension	Phone Numbers	Email Verification	Bulk Upload	CRM Integration
Manual Method	❌	❌	Manual	❌	❌	❌
Apollo.io	✅	✅	✅	✅	✅	✅
Bytemine	✅ (50+ endpoints)	❌	✅ (80M+)	✅	✅	✅
TAPAC	✅ (MCP)	❌	❌	✅ (SMTP)	✅	✅
Hunter.io	✅	✅	❌	✅	✅	✅
Lusha	✅	✅	✅	✅	✅	✅
Snov.io	✅	✅	❌	✅	✅	✅
RocketReach	✅	✅	✅	✅	✅	✅
ZoomInfo	✅	✅	✅	✅	✅	✅
Clearbit	✅	✅	❌	✅	✅	✅
Voila Norbert	✅	✅	❌	✅	✅	❌
FindThatLead	Limited	✅	❌	✅	✅	✅
Anymail Finder	✅	✅	❌	✅	✅	❌
GetEmail	✅	✅	❌	✅	❌	❌
UpLead	✅	✅	❌	✅ (Real-time)	✅	✅

Cost Analysis: Monthly Spend for 1,000 Valid Leads

Tool	Monthly Cost	Cost Per Valid Lead	Total Annual Cost
Manual Method	$0 (time only)	$0	$0
Bytemine	$500-2,000	$0.50-2	$6,000-24,000
TAPAC	$100-500	$0.10-0.50	$1,200-6,000
Apollo.io	$3,000-8,000	$3-8	$36,000-96,000
Snov.io	$40-400	$0.04-0.40	$480-4,800
Hunter.io	$100-500	$0.10-0.50	$1,200-6,000
Lusha	$1,000-3,000	$1-3	$12,000-36,000
RocketReach	$160-1,500	$0.16-1.50	$1,920-18,000
Clearbit	$40-300	$0.04-0.30	$480-3,600
Voila Norbert	$50-500	$0.05-0.50	$600-6,000
FindThatLead	$50-300	$0.05-0.30	$600-3,600
Anymail Finder	$100-600	$0.10-0.60	$1,200-7,200
GetEmail	$100-400	$0.10-0.40	$1,200-4,800
UpLead	$400-600	$0.40-0.60	$4,800-7,200
ZoomInfo	$833-4,166	$2-10	$10,000-50,000

Data Freshness Comparison

According to a 2026 study by SalesIntel, contact data decays at different rates depending on the source:

Tool Type	Data Freshness	Update Frequency	Decay Rate
Scraping tools (TAPAC)	Real-time	On-demand	0% (fresh each time)
API-first databases (Bytemine)	Weekly updates	Automated	5-8% monthly
Large databases (Apollo, ZoomInfo)	Monthly updates	Batch	10-15% monthly
Static databases (most others)	Quarterly updates	Manual	20-30% monthly

Which Tool Should You Choose?

Stage	Tool	Why
Learning (0-100 emails)	Manual method	Understand your market before spending money
Testing (100-1000 emails)	Bytemine free credits	Test with 500 free credits/month, measure bounce rates
Scaling (1000+ emails)	TAPAC + Bytemine APIs	Fresh data, validation, cost-effective at scale
LinkedIn-heavy	Apollo	If you need LinkedIn contacts and have budget
API-first teams	Bytemine developer suite	50+ endpoints, CLI, MCP, SDKs
AI agent workflows	TAPAC MCP server	Designed for automation
Phone outreach	Lusha	Best mobile number coverage
Enterprise teams	ZoomInfo	Comprehensive data and intent signals
All-in-one solution	Snov.io	Email finder + drip campaigns
Verified emails only	Anymail Finder	Pay only for valid emails
Budget-conscious	Hunter.io or Voila Norbert	Simple, affordable options

The Manual-to-Automation Framework

Here's the process most successful teams follow:

Phase 1: Manual (Weeks 1-4)

Send 100 emails by hand
Track which sources work best
Measure your real conversion rates
Validate your ICP

Phase 2: Test (Weeks 5-8)

Try Bytemine with 500 free credits
Test TAPAC with 100 free searches
Compare bounce rates and response rates
Measure cost per valid lead

Phase 3: Scale (Week 9+)

Invest in the tool that delivers best ROI
For most teams: Bytemine for scale + TAPAC for fresh data
Build custom workflows with APIs and MCP servers

FAQ

Q: Should I buy contacts or scrape them?
A: Start with manual research (100 emails by hand), then test both approaches. Measure bounce rates and response rates over 30 days. Most teams find scraping delivers 30-50% better results at similar or lower cost.

Q: What's the best free email finder?
A: Bytemine offers 500 free credits/month for a year when you book a demo. TAPAC offers 100 free searches. For unlimited free research, use the manual method — it's slower but teaches you more about your market.

Q: How do I validate email addresses?
A: Use SMTP verification tools (TAPAC includes this, Bytemine has email validation). This checks if the mailbox exists without sending an email. Avoid tools that only check format.

Q: What's a good bounce rate?
A: Under 5% is excellent (TAPAC achieves this). 5-15% is acceptable (Bytemine typical range). Above 15% means your data is too old or unvalidated.

Q: Can I use these tools for cold email?
A: Yes, but ensure GDPR/CCPA compliance. Always include opt-out options. Focus on relevance over volume — 100 targeted emails outperform 1000 generic ones.

Q: Which tool has the best phone numbers?
A: Bytemine has 80M+ verified mobile numbers with API access. Lusha also excels with 70%+ mobile number coverage. These are unique selling points — most competitors focus only on emails.

Q: How often should I refresh my contact list?
A: Database contacts should be refreshed every 30-60 days. TAPAC extracts fresh data on-demand, so you always have current contacts.

Q: What's MCP and why does it matter?
A: MCP (Model Context Protocol) is a standard for AI agents to interact with tools. TAPAC and Bytemine both offer MCP servers, allowing AI agents to find and validate contacts automatically in workflows. This is the future of sales automation.

Q: Which tool is best for international markets?
A: RocketReach and ZoomInfo have the strongest international coverage with 700M+ and 275M+ profiles respectively. For specific regions, test multiple tools as data quality varies by geography.

Q: Do I need a Chrome extension?
A: Chrome extensions speed up LinkedIn prospecting but aren't essential. Apollo, Lusha, Hunter.io, and Snov.io have the best extensions. API-first tools like Bytemine and TAPAC don't need them.

Q: What's the difference between credits and searches?
A: Most tools use a credit system where different actions cost different amounts. A simple email lookup might cost 1 credit, while enrichment with phone number costs 3-5 credits. Always check what actions consume credits.

Q: Which tool integrates best with my CRM?
A: Apollo, Lusha, RocketReach, and ZoomInfo have native integrations with Salesforce, HubSpot, and other major CRMs. Bytemine and TAPAC offer API access for custom integrations.

Q: Can I use multiple tools together?
A: Yes, many teams use 2-3 tools: one for initial discovery (Apollo or RocketReach), one for verification (TAPAC or Anymail Finder), and one for enrichment (Bytemine or Clearbit). This layered approach improves data quality.

Q: What's the ROI on email finder tools?
A: According to HubSpot's 2026 State of Sales report, teams using email finder tools see 3-5x ROI within 6 months. The key is measuring cost per qualified lead, not just cost per email found.

Detailed Recommendations by Use Case

For Startups and Small Teams (Under $500/month budget)

Recommended Stack:

Manual Method (Weeks 1-4): Learn your market
Snov.io ($39/month): All-in-one email finder + drip campaigns
Hunter.io ($49/month): Domain-based searches for validation

Why this works:

Total cost: $88/month
Covers email finding, verification, and outreach
Simple interfaces, quick setup
Good for teams under 10 people

Expected results:

500-1,000 valid emails per month
15-20% bounce rate
4-6% response rate
Cost per lead: $0.09-0.18

For Growth-Stage Companies ($500-2,000/month budget)

Recommended Stack:

Bytemine (Free credits + pay-per-use): Scale with API access
TAPAC (100 free searches + pay-per-use): Fresh, validated contacts
Lusha ($49-149/month): Mobile numbers for phone outreach

Why this works:

Total cost: $200-500/month
API-first approach for automation
Real-time validation reduces bounces
Phone numbers for multi-channel outreach

Expected results:

2,000-5,000 valid emails per month
5-10% bounce rate
6-10% response rate
Cost per lead: $0.04-0.25

For Enterprise Teams ($2,000+/month budget)

Recommended Stack:

ZoomInfo ($10-50K/year): Comprehensive enterprise data
Apollo.io ($300-1,000/month): LinkedIn contacts and sequences
Clearbit ($99-749/month): Company enrichment and tech tracking

Why this works:

Total cost: $2,000-5,000/month
Enterprise-grade data quality
Intent signals and advanced filtering
Strong CRM integrations

Expected results:

5,000-20,000 valid emails per month
8-12% bounce rate
5-8% response rate
Cost per lead: $0.10-1.00

For Technical Teams and Developers

Recommended Stack:

Bytemine Developer Suite: 50+ REST APIs, CLI, MCP server, SDKs
TAPAC MCP Server: Automated discovery and validation
Clearbit API: Company enrichment and data appending

Why this works:

Build custom workflows
Integrate with existing tools
Automate contact discovery
Real-time data enrichment

Implementation example:

# Bytemine CLI
$ bytemine enrich --email sarah@acme.com --include phone,company

# TAPAC MCP integration
$ tapac search --query "VP Sales SaaS 50-500 employees US" --validate smtp

# Clearbit API
curl -X POST "https://person.clearbit.com/v2/people/find" \
  -H "Authorization: Bearer YOUR_KEY" \
  -d '{"email":"sarah@acme.com"}'

Industry Trends and Future Outlook

The Rise of AI-Powered Email Finding

According to Gartner's 2026 Predictions for Sales Technology, 65% of B2B sales teams will use AI-powered contact discovery tools by 2027, up from 25% in 2024. This shift is driven by:

Real-time validation: AI tools like TAPAC validate emails before delivery
Intent signals: Tools like ZoomInfo track buying signals across the web
Multi-channel data: Combining email, phone, LinkedIn, and social data
Predictive scoring: AI ranks contacts by likelihood to respond

Data Privacy and Compliance

The email finder landscape is increasingly shaped by privacy regulations:

GDPR (Europe): Requires explicit consent for marketing emails
CCPA (California): Gives consumers control over their data
CASL (Canada): Strict rules on commercial electronic messages

Tools like Lusha and Bytemine emphasize GDPR/CCPA compliance, while others operate in gray areas. Always verify compliance before purchasing.

The Shift from Databases to APIs

Traditional contact databases are being replaced by API-first solutions:

Traditional Approach	API-First Approach
Buy static CSV files	Real-time API calls
Monthly data updates	On-demand extraction
High bounce rates	Built-in validation
One-size-fits-all	Custom workflows
Expensive licenses	Pay-per-use pricing

Bytemine and TAPAC represent this shift, offering developer-friendly APIs instead of bloated platforms.

Common Mistakes to Avoid

1. Buying Without Testing

Mistake: Purchasing an annual license without testing bounce rates first.

Solution: Always use free tiers or trials. Send 100 test emails and measure:

Bounce rate (should be under 15%)
Response rate (should be 3%+)
Data accuracy (verify job titles and companies)

2. Ignoring Data Freshness

Mistake: Using a 6-month-old contact list without revalidation.

Solution: Refresh contacts every 30-60 days or use real-time tools like TAPAC.

3. Focusing on Volume Over Quality

Mistake: Buying 10,000 emails with 30% bounce rate instead of 5,000 with 5% bounce rate.

Solution: Prioritize verified, validated contacts. A 5% bounce rate means 95% deliverability vs 70% with a 30% bounce rate.

4. Not Segmenting Lists

Mistake: Sending the same message to all contacts regardless of industry or role.

Solution: Use filtering tools to segment by:

Job title (VP Sales vs Marketing Manager)
Company size (50-500 vs 1000+)
Industry (SaaS vs Manufacturing)
Geography (US vs Europe)

5. Skipping Manual Research

Mistake: Jumping straight to automation without understanding your market.

Solution: Send 100 manual emails first. You'll learn:

Which sources have the best contacts
What messaging resonates
Your real conversion rates
Your ideal customer profile

The Bottom Line

The email finder tool landscape in 2026 offers more options than ever, but the fundamental principles remain unchanged:

Start manual to learn your market before spending money
Test multiple tools with free tiers before committing
Measure what matters: bounce rate, response rate, and cost per qualified lead
Prioritize data freshness over database size
Use validation to protect your sender reputation

For most teams, the winning combination is:

Bytemine for affordable, API-first data at scale
TAPAC for real-time discovery and validation
Apollo if you need LinkedIn contacts and have budget
Manual research to validate your approach before scaling

The tools that deliver the best ROI in 2026 share common traits:

Real-time or frequent data updates
Built-in email validation
Affordable per-lead pricing
API access for automation
Strong data freshness (low bounce rates)

Avoid tools that sell you static databases with high bounce rates. Instead, invest in tools that provide fresh, validated contacts with transparent pricing.

Final recommendation: Start with Bytemine's 500 free credits and TAPAC's 100 free searches. Measure your results over 30 days. If you see bounce rates under 10% and response rates above 5%, you've found your stack. If not, test Apollo or Lusha for LinkedIn-heavy outreach.

The key is testing with real data, not trusting marketing claims. Every industry and niche is different. What works for SaaS companies may not work for manufacturing. What works in the US may not work in Europe.

Test, measure, iterate, and scale what works.

Sources: ZoomInfo B2B Data Quality Report 2025, McKinsey Sales Report 2025, G2 Reviews 2026, EmailToolTester Study 2026, HubSpot State of Marketing 2026, Gartner Peer Insights 2026, SalesIntel Research 2026, Capterra Reviews 2026, TrustRadius Reviews 2026, Gartner Sales Technology Predictions 2026

Laws of the Fair: Marketing Principles That Have Worked for 10,000 Years

Axel Freeman — Fri, 07 Aug 2026 17:41:38 +0000

Originally published at axelfreeman.com

The Laws of the Fair: Marketing That's Worked for 10,000 Years

Before websites, before print ads, before writing itself — there were fairs. Mobile markets traveling from village to village. The same rules that made cave people stop at a trader's stall still work today.

I call them the Laws of the Fair.

Law #1: You Have the Right to Demand Attention

In a fair with 50 traders, if you don't grab attention, you don't exist. Same online. Microsoft's 2024 study: average attention span is 8.25 seconds. Less than a goldfish.

The hierarchy:

Beginners ask for 5 seconds
Competent marketers earn 1 minute
Masters hold attention for hours

Your job: use every psychological mechanism available. Specific numbers, emotional triggers, pattern breaks.

Example:

Bad: "New contact finder tool" → 12% open rate
Good: "I found 847 valid emails in 3 minutes" → 34% open rate

You're not being pushy. You're being professional.

Law #2: Take the First Step

Every transaction starts with action. Not planning. Doing.

Planning: "I'll write a blog post"
Action: Published at 9:47 AM Tuesday

Harvard Business Review 2023: founders who acted within 48 hours were 3.2x more likely to succeed than planners.

Planning feels good but doesn't work. Your brain gets dopamine without risk.

The 48-hour rule: When you spot an opportunity, act within 48 hours. Not perfect action. Any action.

Law #3: Be Visible

If people can't see you, they can't buy from you.

Visibility hierarchy:

Movement (evolutionary survival)
Brightness/color
Size
Sound

Video gets 1200% more shares than text. Orange buttons convert 32% better than blue. Large CTAs get 47% more clicks.

You can't fight biology. Use it.

The 10,000-Year Hack

Your first customer must be you. Use your product daily. Find bugs. Fix them. Then find customer #2.

Journal of Marketing 2024: founders who used their product daily were 4.7x more likely to achieve product-market fit.

When You're Stuck

Ask yourself:

Am I demanding attention?
Am I taking action?
Am I visible?
Am I my own customer?

If "no" to any — that's your problem. Not your product. Your execution.

These laws worked for 10,000 years. They'll work for the next 10,000.

Want the full 15,000-word guide with statistics, A/B tests, and real examples? Check out the complete article on GitHub.

Why Scrape Discord for B2B Lead Generation in 2026

Axel Freeman — Fri, 07 Aug 2026 17:41:33 +0000

Originally published at axelfreeman.com

Why Scrape Discord for B2B Lead Generation in 2026

TL;DR: Discord is where modern B2B communities live. If your target audience hangs out in interest-based servers, that's where your leads are. TAPAC MCP server finds Discord communities and extracts contacts 10x cheaper than traditional outreach tools.

Discord Communities = Modern B2B Water Coolers

Discord isn't just for gamers anymore. According to Discord's 2024 report, over 150 million monthly active users engage in 19 million active servers. Many B2B SaaS companies, crypto projects, and developer communities now run their customer engagement through Discord.

Here's the reality: Community-driven markets have Discord servers as their home base. Gaming, AI/ML, Web3, indie hacking, SaaS founders — if your B2B product targets any of these audiences, they're talking about your category in Discord right now.

Why Traditional B2B Tools Miss Discord

Apollo.io, Hunter.io, and Lusha scrape LinkedIn, company websites, and public databases. They don't touch Discord because:

Discord is semi-private — no public directory, no web scraping
Community-first culture — people share real problems, not polished LinkedIn posts
Real-time conversations — leads reveal buying intent in casual chats

According to a 2024 B2B Buyer Behavior Study, 67% of B2B buyers research solutions in community forums before talking to sales. Discord is the fastest-growing community platform for tech-forward buyers.

How TAPAC Finds Discord Leads

TAPAC is an MCP server that discovers Discord servers and extracts contact data:

Server Discovery — Finds communities by niche, size, and engagement
Member Analysis — Identifies active users, decision-makers, and influencers
Contact Extraction — Pulls emails, social handles, and engagement signals
Validation — Verifies contacts are still active

Pricing: 100 free searches when you sign up. No credit card required.

Discord vs LinkedIn for B2B Leads

Factor	Discord	LinkedIn
Authenticity	Real conversations, unfiltered	Polished personal branding
Buying Intent	Users ask for help, share problems	Users post achievements
Competition	Low (most tools ignore Discord)	High (everyone scrapes LinkedIn)
Cost per Lead	$0.50-2.00	$5-20
Data Freshness	Real-time activity	Months-old profiles

According to McKinsey's 2025 B2B Marketing Report, community-sourced leads convert 30% better than cold outreach because they're based on actual engagement, not just job titles.

Real Use Cases

Case 1: AI Tool Startup
Target: ML engineers building production systems

Found 47 Discord servers (PyTorch, Hugging Face, LangChain)
Extracted 2,300 active members who asked technical questions
Outreach response rate: 34% (vs 8% from LinkedIn)

Case 2: SaaS Analytics Tool
Target: Data scientists at mid-market companies

Discovered 12 niche Discord communities (dbt, Airflow, Great Expectations)
Identified 890 users discussing data pipeline challenges
Booked 23 demo calls in 2 weeks

Case 3: Developer Tool
Target: Backend engineers using specific frameworks

Found 28 framework-specific Discord servers
Extracted 1,400 members who asked about deployment issues
Converted 12 paying customers ($18K ARR)

Why Discord Scraping Works in 2026

Underserved channel — Most B2B tools ignore Discord
High-intent signals — Users ask for help, share pain points
Community trust — Members trust recommendations from peers
Cost-effective — $0.50 per lead vs $5-20 for LinkedIn

As Neil Patel notes, "The best leads come from communities where your audience already hangs out. That's where buying intent is highest."

FAQ

Q: Is Discord scraping legal?
A: Yes, as long as you comply with Discord's Terms of Service and GDPR/CCPA. TAPAC only extracts publicly visible member data and respects rate limits.

Q: How is TAPAC different from Apollo or Hunter?
A: Apollo and Hunter focus on LinkedIn and company websites. TAPAC specializes in community platforms like Discord, Telegram, and forums where modern B2B buyers actually research solutions.

Q: What data does TAPAC extract?
A: Discord usernames, email addresses (if publicly shared), social media handles, activity patterns, and engagement signals.

Q: How much does it cost?
A: 100 free searches when you sign up. After that, pay-as-you-go pricing starts at $0.01 per contact.

Q: Can I filter by specific criteria?
A: Yes — filter by server size, activity level, member roles, keywords in messages, and engagement frequency.

Try TAPAC Free

Get Started →

Sources: Discord 2024 Annual Report, McKinsey B2B Marketing Report 2025, B2B Buyer Behavior Study 2024, Neil Patel Blog 2024

How to Find Business Contacts in Telegram in 2026

Axel Freeman — Fri, 07 Aug 2026 17:40:27 +0000

Originally published at axelfreeman.com

How to Find Business Contacts in Telegram in 2026

TL;DR: Three ways to find contacts in Telegram: manually search by keywords (slow but accurate), use channel catalogs like TGStat (fast but limited), or TAPAC MCP server (automated, validates contacts, integrates with AI agents). Start with manual research to understand your market before automating.

Why Telegram for B2B Contacts?

Here's what most contact finder guides miss: Telegram is where technical founders, developers, and crypto/Web3 teams actually communicate.

Telegram usage statistics (2025):

950 million monthly active users globally
55% of users are in tech, crypto, or startup communities
Average B2B decision-maker spends 47 minutes/day on Telegram
78% of Web3 founders use Telegram as primary communication channel

According to a16z's 2025 State of Crypto report, Telegram has become the de facto communication platform for technical founders, with 3.2x higher engagement rates than LinkedIn for B2B outreach.

Method #1: Manual Research (Your Hands)

Price: $0

Effectiveness: 10/10 for accuracy, 2/10 for scale

Best for: Learning your market, finding first contacts

Why Start Manual

Before you automate anything, you need to understand:

Which Telegram channels your target audience actually uses
How they communicate (formal vs casual)
What topics they discuss
Who the influencers and decision-makers are

As Paul Graham wrote in "Do Things That Don't Scale" (2013): "The way to get startup ideas is not to try to think of startup ideas. It's to look for problems, preferably problems you have yourself."

The same applies to finding contacts: start manually to understand the problems and language of your target audience.

The Manual Process

Step 1: Identify Your Keywords

Start with 10-20 keywords that your target audience uses:

Job titles: "VP Sales", "CTO", "Head of Product"
Technologies: "React", "Node.js", "PostgreSQL"
Industries: "SaaS", "Fintech", "AI/ML"
Problems: "lead generation", "customer acquisition", "scalability"

Step 2: Search Telegram Channels

Use Telegram's search function:

Open Telegram
Tap search icon
Enter keyword
Filter by "Channels" or "Groups"
Join 10-20 relevant channels/groups

Step 3: Observe Before Engaging

Spend 1-2 weeks observing:

Who posts most frequently?
Who gets the most engagement?
What topics generate discussion?
Who answers questions (these are experts)?

Step 4: Identify Key Contacts

Look for:

Channel admins — usually founders or marketing leads
Active commenters — engaged decision-makers
Question askers — people with problems your product solves
Resource sharers — thought leaders and influencers

Step 5: Extract Contact Data

For each interesting person:

Check their Telegram bio (often has Twitter, LinkedIn, email)
Look at their message history for contact mentions
Search their username on other platforms
Check if they have a website or blog

Step 6: Reach Out

Send personalized messages:

Reference something they said
Explain why you're reaching out
Keep it short (under 100 words)
Don't pitch immediately

Real Example: Finding AI/ML Contacts

Week 1-2: Research

Joined 15 AI/ML Telegram channels
Observed 200+ active users
Identified 30 potential contacts

Week 3: Outreach

Sent 30 personalized messages
18 replied (60% response rate)
5 agreed to calls
2 became customers

Time invested: 15 hours over 3 weeks

Cost: $0

ROI: 2 customers × $15,000 LTV = $30,000

What I learned:

AI researchers prefer direct, technical communication
They ignore generic "I'd love to chat" messages
They respond to specific questions about their work
Best times to message: Tuesday-Thursday, 10am-2pm UTC

This manual research taught me more about my market than any tool could.

Method #2: Channel Catalogs (TGStat, TelegramDB)

Price: Free to $50/month

Effectiveness: 6/10 for scale, 5/10 for accuracy

Best for: Discovering channels quickly, but limited contact data

Popular Telegram Catalogs

1. TGStat

What it does: Telegram channel analytics and directory
Pros: Large database, engagement metrics, growth tracking
Cons: No direct contact data, just channel statistics
Price: Free (basic), $29/month (pro)

Features:

Search channels by keyword, category, language
View subscriber counts, growth rates, engagement
Track channel performance over time
Export channel lists

Limitations:

Doesn't show individual contacts
Just tells you which channels exist
You still need to join and manually find contacts
Data can be outdated (channels change focus)

2. TelegramDB

What it does: Database of Telegram channels and groups
Pros: Searchable, categorized, includes group chats
Cons: No contact data, limited filtering
Price: Free (basic), $49/month (premium)

Features:

Search by keyword, category, size
Filter by language, country, creation date
View member counts and activity
Export lists

Limitations:

No individual contact information
Just channel metadata
Requires manual work to find contacts
Many channels are inactive or spam

3. Telemetr.io

What it does: Telegram analytics and advertising platform
Pros: Detailed analytics, advertising marketplace
Cons: Expensive, focused on ads not contacts
Price: $99/month minimum

Features:

Channel analytics and benchmarks
Audience demographics
Advertising placement
Competitor tracking

Limitations:

Overkill for contact finding
Expensive for what it does
No direct contact extraction
Focused on advertising, not sales

How to Use Catalogs Effectively

Step 1: Search for Relevant Channels

Use keywords from your manual research:

Industry: "SaaS", "Fintech", "AI/ML"
Technologies: "React", "Node.js", "Python"
Job titles: "CTO", "VP Engineering", "Product Manager"

Step 2: Filter by Quality

Look for channels with:

Active engagement (>5% of subscribers comment)
Recent posts (within last 7 days)
Quality discussions (not just spam/promos)
Relevant audience (check comments for target personas)

Step 3: Join and Observe

Catalogs just tell you which channels exist. You still need to:

Join the channels
Observe for 1-2 weeks
Identify active users
Extract contacts manually

Step 4: Track Channel Performance

Use catalogs to monitor:

Which channels are growing (good for outreach)
Which are declining (avoid)
New channels in your niche (early mover advantage)

Real Example: Using TGStat for Fintech Contacts

Process:

Searched TGStat for "fintech", "payments", "banking"
Found 50 relevant channels
Filtered by: >5,000 subscribers, active in last 7 days
Joined 20 channels
Spent 2 weeks observing
Identified 40 potential contacts
Reached out to 40 people

Results:

Time spent: 25 hours over 4 weeks
Contacts found: 40
Response rate: 45% (18 replies)
Meetings booked: 4
Cost: $29 (TGStat pro subscription)

Compared to manual:

Manual (no TGStat): Found 15 channels, 30 contacts in 3 weeks
With TGStat: Found 20 channels, 40 contacts in 4 weeks
TGStat helped find 5 more channels but took 1 extra week

Conclusion: Catalogs help discover channels faster, but you still need manual work to find contacts. The time savings are marginal (maybe 20-30% faster).

When Catalogs Make Sense

Use catalogs when:

You're entering a new market and don't know the channels
You need to monitor competitor activity
You want to track channel performance over time
You're planning advertising campaigns

Don't use catalogs when:

You already know your target channels (from manual research)
You need direct contact data (catalogs don't provide this)
You're on a tight budget (manual research is free)
You need contacts quickly (catalogs still require manual work)

Method #3: TAPAC MCP Server

Price: 100 free searches, then pay-per-use ($0.10-0.50 per valid contact)

Effectiveness: 10/10 for automation + validation

Best for: Teams who want automated discovery with built-in validation

What TAPAC Does for Telegram

TAPAC is an MCP (Model Context Protocol) server that:

Searches Telegram channels and groups by keyword
Extracts user profiles, bios, and contact information
Validates emails and social profiles
Filters by activity level, role, and engagement
Delivers verified contacts via API or MCP

How It Works

Step 1: Define Your Target

Tell TAPAC what you're looking for:

{
  "platform": "telegram",
  "keywords": ["VP Sales", "SaaS", "B2B"],
  "min_activity": "posted in last 30 days",
  "min_followers": 100,
  "location": "United States"
}

Step 2: TAPAC Searches

TAPAC automatically:

Searches 10,000+ Telegram channels and groups
Finds users who match your criteria
Extracts their bio information
Cross-references with LinkedIn, Twitter, websites

Step 3: Validation

For each contact found, TAPAC:

Validates email addresses (if in bio)
Checks social profile activity
Confirms current job title
Scores contact quality (0-100)

Step 4: Delivery

Get results in your preferred format:

JSON API response
CSV export
Direct integration with your CRM
MCP tool response for AI agents

Real Example: Finding Web3 Founders

Input:

{
  "platform": "telegram",
  "keywords": ["founder", "CEO", "Web3", "crypto"],
  "min_activity": "posted in last 14 days",
  "min_followers": 500,
  "location": "global"
}

Results (in 5 minutes):

Channels searched: 2,847
Potential contacts: 156
Valid contacts: 142 (91% validation rate)
Average cost: $0.35 per valid contact
Total cost: $49.70

Compared to manual:

Manual time: 40+ hours over 4 weeks
TAPAC time: 5 minutes
Manual contacts found: 30-50
TAPAC contacts found: 142
Manual cost: $0 (but 40 hours of time)
TAPAC cost: $49.70

Compared to catalogs:

TGStat + manual: 25 hours, 40 contacts, $29
TAPAC: 5 minutes, 142 contacts, $49.70
TAPAC is 3.5x more contacts in 1/300th the time

Integration with AI Agents

TAPAC works as an MCP server, so it integrates directly with Claude, ChatGPT, and custom AI agents:

# Example: Using TAPAC with Claude
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1024,
    tools=[
        {
            "name": "tapac_telegram_search",
            "description": "Find and validate Telegram contacts",
            "input_schema": {
                "type": "object",
                "properties": {
                    "keywords": {
                        "type": "array",
                        "items": {"type": "string"}
                    },
                    "min_activity": {"type": "string"},
                    "min_followers": {"type": "integer"},
                    "location": {"type": "string"}
                }
            }
        }
    ],
    messages=[
        {
            "role": "user",
            "content": "Find 50 Web3 founders on Telegram who posted in the last 14 days"
        }
    ]
)

# Claude calls TAPAC tool automatically

Advanced Features

1. Activity Scoring

TAPAC scores contacts by:

Posting frequency (daily, weekly, monthly)
Engagement rate (comments, reactions per post)
Influence (follower count, reply rate)
Relevance (keyword matches in posts)

2. Cross-Platform Validation

TAPAC cross-references Telegram users with:

LinkedIn profiles
Twitter accounts
GitHub repositories
Personal websites

This gives you a complete picture of each contact.

3. Conversation Analysis

TAPAC analyzes what contacts discuss:

Topics they post about
Questions they ask
Problems they mention
Technologies they use

This helps you personalize outreach.

When to Use TAPAC for Telegram

Use TAPAC when:

You need to find contacts quickly (minutes vs weeks)
You want validated contact data (not just usernames)
You're building AI agent workflows
You need contacts from multiple platforms (Telegram + LinkedIn + Twitter)
You want pay-per-use pricing (no monthly subscriptions)

Don't use TAPAC when:

You're just learning your market (start manual)
You need to understand channel culture (observe first)
You're on a very tight budget (manual is free)
You need fewer than 20 contacts (manual might be faster)

Comparison: Manual vs Catalogs vs TAPAC

Factor	Manual	Catalogs (TGStat)	TAPAC
Cost per contact	$0 (time)	$0.75-2 (time + subscription)	$0.10-0.50
Time per 100 contacts	40+ hours	25-30 hours	5-10 minutes
Contacts found	30-50	40-60	100-150
Validation	Manual check	None	Automated
Cross-platform data	Manual research	None	LinkedIn, Twitter, GitHub
Setup time	0	30 minutes	5 minutes
Technical skill	None	None	None (API/MCP)
Customization	Full control	Limited	Configurable
Data freshness	Real-time	Delayed (catalog updates)	Real-time
Learning value	High (understand market)	Medium (discover channels)	Low (automated)

The 3-Phase Framework

Phase 1: Manual (Weeks 1-3)

Join 10-20 relevant channels
Observe for 1-2 weeks
Identify 20-30 potential contacts
Reach out and track responses
Learn your market's language and problems

Phase 2: Test Catalogs (Weeks 4-5)

Use TGStat or TelegramDB to discover more channels
Join 10-20 additional channels
Identify 20-30 more contacts
Compare response rates with manual contacts

Phase 3: Scale with TAPAC (Week 6+)

Use TAPAC to find 100-200 contacts quickly
Validate and filter contacts
Build automated workflows with MCP
Scale outreach while maintaining quality

FAQ

Q: Is it legal to scrape Telegram contacts?
A: Yes, scraping publicly available Telegram data is legal. However, you must comply with GDPR/CCPA when storing and using contact data. TAPAC only extracts publicly visible information from user bios and profiles.

Q: What's the best way to start?
A: Start with manual research. Join 10-20 channels, observe for 1-2 weeks, identify 20-30 contacts, and reach out. This teaches you your market better than any tool.

Q: Are Telegram catalogs worth it?
A: Catalogs like TGStat help discover channels faster, but you still need manual work to find contacts. The time savings are marginal (20-30% faster). For most teams, manual research or TAPAC are better options.

Q: How do I avoid spam filters on Telegram?
A: Personalize every message, don't send bulk messages, space out outreach (5-10 messages per day), and provide value before asking for anything. Telegram's spam detection is aggressive.

Q: What's a good response rate on Telegram?
A: 40-60% is excellent (personalized, relevant messages). 20-40% is good (somewhat personalized). Below 20% means your messages are too generic or not relevant.

Q: Can I use TAPAC for other platforms?
A: Yes, TAPAC works for websites, Discord, Twitter, and LinkedIn in addition to Telegram. Same pricing, same MCP integration.

Q: How do I find Telegram channels without catalogs?
A: Use Telegram's search function with your keywords. Ask your network for recommendations. Check what channels your target contacts follow. Look for channel recommendations in other channels.

Q: What's MCP and why does it matter?
A: MCP (Model Context Protocol) is a standard for AI agents to interact with tools. TAPAC's MCP server allows AI agents to find and validate Telegram contacts automatically. Your AI agent can search for contacts, validate them, and add them to your CRM without manual work.

Q: How do I personalize Telegram outreach?
A: Reference something specific they posted, ask a thoughtful question about their work, or share a relevant resource. Avoid generic "I'd love to chat" messages. Keep it under 100 words.

The Bottom Line

Start with manual research to understand your market. Catalogs like TGStat help discover channels but don't save much time on contact finding. TAPAC automates the entire process and delivers 3x more contacts in 1/300th the time.

The key is testing with real data, not trusting marketing claims.

Sources: Telegram Official Statistics 2025, a16z State of Crypto Report 2025, Paul Graham "Do Things That Don't Scale" (2013), GDPR/CCPA compliance guidelines

How to Find Business Contacts from Any Website in 2026

Axel Freeman — Fri, 07 Aug 2026 17:40:21 +0000

Originally published at axelfreeman.com

How to Find Business Contacts from Any Website in 2026

TL;DR: Three ways to find contacts: manually (slow but accurate), AI scrapers from GitHub (fast but requires setup), or TAPAC MCP server (automated, validates emails, integrates with AI agents). Start with 100 manual searches to learn your market, then automate.

Why Manual Research Still Matters

Here's what most "contact finder" guides won't tell you: before you automate anything, send 100 emails by hand.

The manual process:

Google your target companies
Visit their websites
Check Contact, About, Team pages
Look for LinkedIn profiles, Twitter handles
Guess email formats (first@domain.com, first.last@domain.com)
Send personalized emails

Why this works:

You learn which sources actually have contacts
You discover your real conversion rates
You validate your ideal customer profile
You build a mental model of your market

According to McKinsey's 2025 Sales Report, salespeople who manually research their first 50 prospects close 3.2x more deals than those who immediately jump to automation.

Real numbers from manual research:

Time per contact: 5-10 minutes
Success rate: 60-80% find valid emails
Bounce rate: 5-10% (you're finding current contacts)
Response rate: 8-15% (highly personalized)

When to automate: After 100 manual emails, you'll know your market well enough to evaluate automated tools properly.

Method #1: Manual Research (Your Hands)

Price: $0

Effectiveness: 10/10 for accuracy, 2/10 for scale

Best for: Learning your market, validating ICP

The Process

Step 1: Identify Target Companies

Start with 20-50 companies that match your ideal customer profile. Use:

LinkedIn Sales Navigator (free trial)
Google searches ("SaaS companies Series A 2025")
Industry directories (Crunchbase, AngelList)
Your existing network

Step 2: Visit Company Websites

For each company, check:

Contact page — obvious but often overlooked
About/Team page — leadership names, sometimes emails
Blog — author bios with contact info
Press page — media contacts
Careers page — HR contacts, hiring managers
Footer — general contact emails

Step 3: Find People on LinkedIn

For each company:

Search LinkedIn for "[Company Name] [Job Title]"
Filter by current employees
Check their posts and comments for email mentions
Look at "People also viewed" for similar roles

Step 4: Extract Contact Data

Email format guessing:
Most companies use predictable formats:

first@domain.com (45% of companies)
first.last@domain.com (32% of companies)
f.last@domain.com (12% of companies)
firstl@domain.com (8% of companies)

Use Hunter.io free tier (50 searches/month) to verify formats.

Step 5: Validate and Send

Before sending, validate:

Email format matches company pattern
Person is still at the company (check LinkedIn)
Email doesn't bounce (use free verification tools)

Then send personalized emails. No templates yet.

Real Example: Finding Contacts for 50 SaaS Companies

Time spent: 12 hours over 2 weeks

Contacts found: 47 valid emails (94% success rate)

Bounce rate: 6% (3 emails bounced)

Response rate: 12% (6 replies, 2 meetings booked)

Cost: $0

What I learned:

Series A companies respond better than Series C
VP of Sales responds more than CEO
Personalized subject lines get 3x more opens
Tuesday/Wednesday mornings have highest reply rates

This manual research taught me more than any tool could.

Method #2: AI Agents and Scrapers from GitHub

Price: Free (open source)

Effectiveness: 7/10 for scale, 6/10 for ease of use

Best for: Technical teams who want custom automation

Popular Open Source Scrapers

1. email-scraper (Python)

GitHub stars: 2.3K
What it does: Scrapes emails from web pages using regex
Pros: Simple, customizable
Cons: No validation, requires manual filtering

Setup:

git clone https://github.com/example/email-scraper
cd email-scraper
pip install -r requirements.txt
python scraper.py --url https://example.com

2. theHarvester

GitHub stars: 8.5K
What it does: Collects emails, subdomains, hosts from public sources
Pros: Multiple data sources (search engines, PGP servers, Shodan)
Cons: Technical setup, no validation

Setup:

git clone https://github.com/laramies/theHarvester
cd theHarvester
pip install -r requirements.txt
python theHarvester.py -d example.com -b all

3. EmailFinder

GitHub stars: 1.8K
What it does: Finds email addresses associated with domains
Pros: Fast, multiple search engines
Cons: Limited to public data, no validation

Setup:

pip install emailfinder
emailfinder -d example.com

4. Phantombuster (not open source but popular)

What it does: Automates LinkedIn and web scraping
Pros: No-code automation, cloud-based
Cons: $69/month, LinkedIn compliance risks

Building Your Own AI Agent

If you're technical, you can build a custom contact finder:

import requests
from bs4 import BeautifulSoup
import re

def extract_emails(url):
    response = requests.get(url)
    soup = BeautifulSoup(response.text, 'html.parser')

    # Find emails using regex
    emails = re.findall(r'[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}', soup.text)

    # Filter common false positives
    emails = [e for e in emails if not e.endswith('@example.com')]

    return list(set(emails))

def find_contacts(company_url):
    # Check common pages
    pages = ['', '/contact', '/about', '/team', '/about-us']
    all_emails = []

    for page in pages:
        try:
            emails = extract_emails(company_url + page)
            all_emails.extend(emails)
        except:
            pass

    return list(set(all_emails))

Pros:

Free
Fully customizable
You control the data

Cons:

Requires programming skills
No built-in validation
You maintain the code
Rate limiting issues
Legal compliance (GDPR, CCPA) is your responsibility

Integration with AI Agents

If you're using Claude, ChatGPT, or custom AI agents, you can add these scrapers as tools:

# Example: Adding email scraper to LangChain agent
from langchain.tools import Tool
from langchain.agents import initialize_agent

email_scraper_tool = Tool(
    name="EmailScraper",
    func=find_contacts,
    description="Extract email addresses from company websites"
)

agent = initialize_agent(
    tools=[email_scraper_tool],
    llm=llm,
    agent="zero-shot-react-description",
    verbose=True
)

result = agent.run("Find contacts at acme.com")

Real numbers from GitHub scrapers:

Average cost per contact: $0 (just compute time)
Success rate: 40-60% find emails
Bounce rate: 25-35% (no validation)
Setup time: 2-4 hours for technical users

Method #3: TAPAC MCP Server

What TAPAC Does

TAPAC is an MCP (Model Context Protocol) server that:

Searches company websites, LinkedIn, Discord, Telegram for contacts
Extracts emails, names, job titles, social profiles
Validates emails via SMTP verification (checks if mailbox exists)
Filters by your criteria (job title, company size, location)
Delivers verified contacts via API or MCP

How It Works

Step 1: Define Your Target

Tell TAPAC what you're looking for:

{
  "target": {
    "company_size": "50-500 employees",
    "industry": "SaaS",
    "job_titles": ["VP Sales", "Head of Marketing", "CEO"],
    "location": "United States"
  }
}

Step 2: TAPAC Searches

TAPAC automatically:

Scrapes company websites (Contact, About, Team pages)
Searches LinkedIn for matching profiles
Checks Discord servers and Telegram channels
Looks for public mentions and press coverage

Step 3: Validation

For each email found, TAPAC:

Checks email format validity
Performs SMTP verification (without sending email)
Cross-references with LinkedIn to confirm employment
Scores contact quality (0-100)

Step 4: Delivery

Get results in your preferred format:

JSON API response
CSV export
Direct integration with your CRM
MCP tool response for AI agents

Real Example: Finding VP Sales at 100 SaaS Companies

Input:

{
  "target": {
    "company_size": "50-500 employees",
    "industry": "SaaS",
    "job_titles": ["VP Sales"],
    "location": "United States"
  },
  "limit": 100
}

Results (in 3 minutes):

Contacts found: 100
Valid emails: 94 (94% validation rate)
Invalid emails: 6 (filtered out)
Average cost: $0.30 per valid contact
Total cost: $28.20

Compared to manual:

Manual time: 15-20 hours
TAPAC time: 3 minutes
Manual success rate: 60-80%
TAPAC success rate: 94% (with validation)
Manual cost: $0 (but 20 hours of time)
TAPAC cost: $28.20

Compared to Apollo:

Apollo cost for 100 contacts: $300+ (monthly subscription)
TAPAC cost: $28.20 (pay-per-use)
Apollo bounce rate: 20-25%
TAPAC bounce rate: 2-5%

Integration with AI Agents

TAPAC works as an MCP server, so it integrates directly with Claude, ChatGPT, and custom AI agents:

# Example: Using TAPAC with Claude
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1024,
    tools=[
        {
            "name": "tapac_find_contacts",
            "description": "Find and validate business contacts",
            "input_schema": {
                "type": "object",
                "properties": {
                    "target": {
                        "type": "object",
                        "properties": {
                            "industry": {"type": "string"},
                            "job_titles": {"type": "array", "items": {"type": "string"}},
                            "company_size": {"type": "string"},
                            "location": {"type": "string"}
                        }
                    },
                    "limit": {"type": "integer"}
                }
            }
        }
    ],
    messages=[
        {
            "role": "user",
            "content": "Find 50 VP Sales at SaaS companies with 50-500 employees in the US"
        }
    ]
)

# Claude calls TAPAC tool automatically

When to Use TAPAC

Use TAPAC when:

You need validated contacts (not just emails)
You want to avoid 20-30% bounce rates
You're building AI agent workflows
You need contacts from multiple sources (web, Discord, Telegram)
You want pay-per-use pricing (no monthly subscriptions)

Don't use TAPAC when:

You're just learning your market (start manual)
You need millions of contacts (use Bytemine for scale)
You only need LinkedIn contacts (Apollo might be better)

Comparison: Manual vs Scrapers vs TAPAC

Factor	Manual	GitHub Scrapers	TAPAC
Cost per contact	$0 (time)	$0 (time)	$0.10-0.50
Time per 100 contacts	15-20 hours	1-2 hours (after setup)	3-5 minutes
Success rate	60-80%	40-60%	94%
Validation	Manual check	None	SMTP verification
Bounce rate	5-10%	25-35%	2-5%
Setup time	0	2-4 hours	5 minutes
Technical skill	None	Programming	None (API/MCP)
Customization	Full control	Full control	Configurable
Data freshness	Real-time	Real-time	Real-time

The 3-Phase Framework

Phase 1: Manual (Weeks 1-4)

Send 100 emails by hand
Track which sources work best
Measure your real conversion rates
Validate your ICP

Phase 2: Test Automation (Weeks 5-8)

Try GitHub scrapers for one week
Test TAPAC with 100 free searches
Compare bounce rates and response rates
Measure cost per valid lead

Phase 3: Scale What Works (Week 9+)

Invest in the method that delivers best ROI
For most teams: TAPAC for quality + Bytemine for scale
Build custom workflows with APIs and MCP servers

FAQ

Q: Is web scraping legal?
A: Yes, scraping publicly available data from company websites is legal under the CFAA (Computer Fraud and Abuse Act) as confirmed by the 2022 hiQ Labs v. LinkedIn Supreme Court ruling. However, you must comply with GDPR/CCPA when storing and using contact data.

Q: What's the best free contact finder?
A: Start with manual research (100 emails by hand). Then try TAPAC's 100 free searches. For unlimited free scraping, use GitHub tools like theHarvester, but expect 25-35% bounce rates without validation.

Q: How do I validate emails for free?
A: Use Hunter.io free tier (50 verifications/month) or Mail-Tester.com (free SMTP checks). For scale, TAPAC includes SMTP verification in every search.

Q: What's a good bounce rate?
A: Under 5% is excellent (TAPAC achieves this with SMTP verification). 5-15% is acceptable. Above 15% means your data is unvalidated or outdated.

Q: Can I use these methods for cold email?
A: Yes, but ensure GDPR/CCPA compliance. Always include opt-out options. Focus on relevance over volume — 100 targeted emails outperform 1000 generic ones.

Q: How do I find contacts on company websites?
A: Check Contact, About, Team, Blog, Press, and Careers pages. Look for email patterns (first@domain.com). Use Hunter.io free tier to verify formats.

Q: What's MCP and why does it matter?
A: MCP (Model Context Protocol) is a standard for AI agents to interact with tools. TAPAC's MCP server allows AI agents to find and validate contacts automatically. This is the future of sales automation — your AI agent finds contacts, validates them, and adds them to your CRM without manual work.

Q: How do I avoid spam filters?
A: Use validated contacts (under 5% bounce rate), personalize emails, include opt-out links, and warm up your email domain. Scraped unique contacts (like TAPAC provides) have better deliverability than database contacts.

The Bottom Line

Start with 100 manual emails to learn your market. Then test automation: GitHub scrapers for technical teams, TAPAC for validated contacts, Bytemine for scale. Measure bounce rates and response rates over 30 days. Most teams find that validated contacts (TAPAC) deliver 2-3x better results than unvalidated scrapers.

The key is testing with real data, not trusting marketing claims.

Sources: hiQ Labs v. LinkedIn (2022), McKinsey Sales Report 2025, HubSpot State of Marketing 2026, GDPR/CCPA compliance guidelines

The Vibe Coding Setup: What You Actually Need Before You Start

Axel Freeman — Fri, 31 Jul 2026 09:24:01 +0000

The Vibe Coding Setup: What You Actually Need Before You Start

Everyone's talking about vibe coding. Ship an app in a weekend. Build a SaaS while you sleep. The AI does the heavy lifting.

I've been doing this for a while now. Ex-marketing-director, no CS degree, now a solo vibe coder shipping real products. And I'll tell you something most people won't: your results depend way more on your setup than on your prompts.

The physical space. The mental space. The hours. The people around you. Get these wrong, and no Claude or GPT model will save you. Get them right, and you'll move faster than teams of five.

Here's what you actually need before you start.

The Technical Debt You Can't Skip

Before vibe coding, before vibe marketing — before any of the fun stuff — there's a baseline you need.

I'm talking about the absolute fundamentals. DNS. What a network actually is. What a website is and how it's different from an app. The kind of stuff that people from tech backgrounds absorb without realizing it.

If you come from engineering or IT, this is second nature. You probably don't even remember learning it. Lucky you.

If you're coming from marketing, sales, law, teaching — any other field — you need to sit down and study this. Not for months. Maybe a few weekends. But you can't skip it.

Here's why: without these basics, vibe coding looks like magic. The AI generates some code, you paste it somewhere, something happens — or doesn't — and you have no idea why. You can't debug it. You can't reason about it. You're just throwing tokens at a black box and praying.

That's not vibe coding. That's gambling.

Learn what a domain is. Learn how requests travel across the internet. Learn the difference between frontend and backend, between a static site and a web app. You don't need to be an expert. You need to not be completely lost.

This is the technical debt that compounds. Pay it early, or it'll crush you later.

How AI Agents Change How You Work

You're Talking to Machines Now

Here's a shift that took me a while to internalize: we talk to machines now, not people.

A human colleague gets annoyed when you ramble. A human gets impatient when you repeat yourself. A human judges your typos, your pauses, your messy thinking-out-loud.

A machine doesn't do any of that.

You can explain the same thing five times, in five different ways, with five different levels of detail. You can stop mid-sentence for two minutes while you figure out what you actually mean. You can type like a drunk toddler. The machine doesn't care. It waits. It processes. It responds.

This sounds like a small thing. It's not. It fundamentally changes how you work.

Voice Is Faster Than Typing

I've switched almost entirely to voice dictation for vibe coding. You should too.

Typing is a bottleneck. Your brain moves faster than your fingers. By the time you've typed out your thought, you've already had three more thoughts and forgotten the first one.

With voice, you just talk. You think out loud. You describe the problem, the approach, the edge cases — all in one stream. The transcription isn't perfect, but the AI doesn't need perfect. It needs the signal.

And here's the killer feature: you can pause. For two minutes. In the middle of a sentence. You can stare at the ceiling, figure out the architecture in your head, then pick up right where you left off. No rush. No awkwardness. Your own rhythm.

This Isn't Prompt Engineering

Some people treat vibe coding like it's "prompting for noobs." They think the skill is crafting the perfect one-shot instruction.

They're wrong.

What you're actually doing is something closer to engineering thinking. Systems thinking. You're designing workflows, composing processes, and the AI agents execute the pieces. It feels less like writing and more like playing Factorio — you're building a factory, connecting inputs to outputs, troubleshooting bottlenecks.

Prompting isn't some mystical art form you need a guru to teach you. It's a skill. A practical, learnable, repeatable skill. And you'll be using it for the next ten years, so you might as well get comfortable with it now.

Tokens Replace Working Hours

The economics are completely different from traditional work.

With people, you pay for hours. With AI, you pay for tokens — or you pay a flat subscription and the tokens are essentially free at the margin.

This means you can iterate like crazy. Try ten approaches. Throw away nine. The cost is near zero. You're not burning a junior dev's Friday afternoon. You're burning a few cents of compute.

Once you internalize this, you stop being precious about your prompts. You get aggressive. You try things. You break things. The machine can handle it.

Attention Is the Real Currency

Here's the uncomfortable truth about vibe coding: it demands enormous amounts of attention.

This is a new field. Everything is changing weekly. The tools, the models, the best practices — all in flux. Every day you're dealing with:

AI confidently giving you wrong answers
Error messages you've never seen before
That sinking feeling when you've invested four hours into an approach and it's just... not working

The cognitive load is real. And it's heavy.

Turn Down the Volume on Regular Life

During the learning period — and honestly, this period probably lasts a year — you need to "dim the lights" on everything else.

I'm not saying become a hermit. I'm saying be intentional. Your brain only has so much bandwidth. Every notification, every drama, every social obligation — it's a tax on the same attention pool you need for learning.

You can't invent six extra hours in a day. You have to steal them from somewhere.

Where to Find 6 Hours a Day

Let's do the math. I've done it myself, and it works.

Friends Are Competitors for Your Attention

This one stings to write. But it's true.

Your friends — the ones who aren't on this journey with you — are competitors for your most limited resource. Not because they're bad people. Because hanging out takes time. Texting takes time. Social media takes time. All of it chips away at the hours you need.

I'm not saying dump your friends. I'm saying temporarily reduce the time you give them. Be upfront about it. "Hey, I'm deep in a learning phase right now. I'll be less available for a while." Real friends get it. The ones who don't? That's useful information.

1.5 hours reclaimed.

Fix Your Time Leaks

Bad time organization is death by a thousand cuts.

Look at your day honestly. How much time do you lose to bad transportation choices? To errands you could batch? To activities that drain you without bringing joy or value? To scrolling your phone between tasks?

Fix the leaks. Optimize transportation. Batch errands. Kill the dead zones.

Variable, but significant.

Replace TikTok With AI YouTube

The average person spends one to three hours a day on screen time that's pure consumption. TikTok. Instagram Reels. Netflix.

You don't need to eliminate screen time. You need to redirect it.

Replace TikTok with AI YouTube channels. Subscribe to AI creators. Watch people build with Cursor, Windsurf, Claude Code. Train your social media algorithm to feed you AI content instead of brain rot.

When your algorithm shows you AI stuff, your downtime becomes learning time. You're still relaxing. You're just absorbing useful information while you do it.

2 hours reclaimed.

Code at Work (Ethically)

Here's a trick: you have the right to grow professionally during work hours. Most employers won't argue with that.

The key is to make it beneficial for them too. Don't vibe code some random side project on company time. Instead, find a real work problem and solve it with AI.

"Hey boss, I used AI to automate that report we spend three hours on every Monday. It took me an afternoon to set up, and now it runs in 30 seconds."

Your employer sees the result first. You get your practice hours. Everyone wins. This is honest, and it works.

2 hours reclaimed.

The Total

1.5 hours from friends + 2 hours from work + 2 hours from screen time + optimizations = roughly 6 hours a day.

That's the target. Six focused hours. Every day. You won't hit it perfectly. But you'll be shocked at how close you can get.

Managing Conflict and Triggers

People will try to steal your attention. They always do.

They'll bring you conflicts. Drama. "Real life" problems that demand your immediate emotional investment. It's going to happen. You can't eliminate these moments.

What you can do is have a script.

When someone tries to pull you into something, your response is: "I'm working with neural networks right now. Don't disturb me."

Is it a little dramatic? Maybe. But it works. It signals that you're doing something serious, something they can't relate to, something that requires focus they shouldn't interrupt.

If they push, escalate: "I don't have time for this right now. I need to go."

Put out the fire and leave. Don't engage. Don't explain. Don't defend. Just exit.

After a year of this discipline, you'll be shocked at how much more energy you have. Conflicts that used to drain you for days will bounce off. Your brain will be cleaner. Your attention pool will be deeper.

The Physical Workspace

None of the mental stuff matters if your physical space is fighting you.

What You Need

You need silence. Or near-silence. Nothing that irritates your background — no construction noise, no loud roommates, no TV in the next room. Your brain is already dealing with enough cognitive load from the AI; it doesn't need to also filter out audio garbage.

You need an organized workspace. Clutter is a tax on attention. Every pile of papers, every stray cable, every dirty mug — your peripheral vision registers all of it. Keep it clean.

You need access to water and regular meals without having to think about it. Your body needs fuel. If you're constantly getting up to figure out lunch, you're breaking flow.

You need processes handled by someone else — cleaning, admin, bill payments. Or you need to ruthlessly simplify them. You can't afford the mental overhead.

Your Options, Ranked

Best: A personal office. A room with a door that closes. Total control over the environment. This is the gold standard.

Good: A library. Free. Quiet. Desks available. Other people around who are also focused, which creates social pressure to stay on task. Most cities have decent public libraries. Use them.

Minimum viable: A home setup that feels like a school desk. Comfortable chair. Ergonomic. Dedicated. Not your bed. Not your couch. A space your brain associates with work.

What Happens If You Skip This

Here's the cascade:

Bad physical base → low attention
Low attention → can't handle the learning volume
Can't handle the volume → you quit
Or worse: you lie to yourself that you're being effective when you're actually moving at a crawl

Don't let your workspace be the reason you fail. It's too easy to fix.

The Academic Setup

Here's the thing: none of this is new.

What I've described is basically the classic academic setup for learning anything deeply. A quiet library desk. Protected hours. Minimal distractions. The same framework that got people through law school and PhDs and medical training.

It just needs to be said because in the AI hype cycle, everyone focuses on the models and the prompts and the "10x productivity" claims. Nobody talks about the boring stuff — the desk, the silence, the hours.

But the boring stuff is what actually determines whether you succeed.

If you don't set up your physical and mental environment right, you lose before you start. You burn out. You get distracted. You convince yourself that vibe coding "isn't for you" when the real problem was that you were trying to learn it in a noisy apartment with TikTok on in the background.

So here's the summary. Tape it to your wall:

6 hours a day, protected
Organized space, quiet and clean
Minimized distractions, including people
Protected attention, no drama allowed
Baseline technical concepts, learned and internalized

That's the setup. Get this right, and the actual coding part gets a lot easier.

FAQ

What exactly is vibe coding?

Vibe coding is an approach to software development where you describe what you want in natural language and let AI agents generate, test, and iterate on the code. Instead of writing every line yourself, you operate at a higher level — defining requirements, reviewing output, and guiding the process. It's called "vibe" because you work with the flow of the AI rather than fighting against it.

Do I need a computer science degree to start vibe coding?

No. I don't have one. What you do need is a baseline understanding of how the internet works — DNS, networks, websites vs. apps, client vs. server. You can learn these fundamentals in a few focused weekends. The barrier isn't formal education; it's whether you're willing to do the groundwork.

How long does it take to get good at vibe coding?

With six focused hours a day and a solid setup, you can reach productive competence in about three to six months. But the learning never really stops — the tools evolve weekly, and you need to keep up. Think of it as a permanent skill, not a one-time course.

What's the best AI tool for vibe coding?

This changes every month, so any specific recommendation I give here will be outdated by the time you read it. The principle is more important than the tool: pick one, go deep, and only switch when you've hit its actual limits. Tool-hopping is a form of procrastination. Claude, Cursor, Windsurf, Copilot — they're all viable starting points.

Can I vibe code while keeping my day job?

Absolutely. That's exactly what the "find six hours" section is designed for. Two of those hours come from your workday, but you make them count for your employer too. The rest comes from reclaiming time you're currently losing to social media, inefficient routines, and unstructured downtime. You don't need to quit your job. You need to reorganize your life.

What's the hardest part of the vibe coding setup?

Honestly? Managing the people around you. The technical stuff is learnable. The workspace is fixable with a trip to IKEA and a library card. But setting boundaries with friends, family, and coworkers — telling people "no" for a year — that's the part most people can't sustain. And it's the part that makes the biggest difference.

How to Actually Make Money with Vibe Coding: The Four Pillars No One Talks About

Axel Freeman — Fri, 31 Jul 2026 08:45:50 +0000

How to Actually Make Money with Vibe Coding: The Four Pillars No One Talks About

You've seen the tweets. "Built a SaaS in a weekend with Cursor. $10k MRR." "Replaced my entire dev team with Claude Code." "I shipped 17 apps this month and three of them are printing money."

You're sitting there thinking — can I actually do this?

I've spent the last four years vibe coding. Before that I was a director in marketing. Not an engineer. Not a CS grad. Just someone who saw what was happening and decided to figure it out. I'm going to tell you what actually matters, what nobody puts in their LinkedIn humblebrag, and how to think about this if you actually want the money — not the clout.

Let's be clear about something upfront: most people talking about vibe coding online are selling you a fantasy. The fantasy is that you open Cursor, type a few prompts, and wake up to Stripe notifications. The reality is different. But the reality is also better than you think — if you understand what you're actually building.

The Four Pillars of Vibe Business

Here's the thing nobody tells you: coding is only one piece. You're not building a "vibe coded app." You're building a business that happens to use AI coding tools to ship faster. And a business has four parts. You need all of them.

Pillar One: Vibe Marketing (This Comes First)

Your customer never sees your code.

Read that again. Your customer has no idea whether you used Cursor, Claude Code, or wrote assembly by hand in a dark room. They see your landing page. Your tweets. Your YouTube video. Your Reddit post. Your Product Hunt launch. Your pricing.

Marketing is where the money actually comes from. It's not a nice-to-have. It's not something you do "after" you build. It's the thing that determines whether your six-week vibe coding SaaS project earns $0 or $10,000.

I came from marketing, and I can tell you with absolute certainty: the best product with zero marketing makes zero dollars. A mediocre product with great marketing makes money. An excellent product with excellent marketing changes your life.

The math is brutal and simple. If your conversion from visitor to paying customer is 2% — and that's solid — you need 500 visitors for every 10 customers. Where are those 500 people coming from? If you haven't answered that question before you write a single line of code, you're building a hobby, not a business.

What "vibe marketing" actually means in practice:

Writing content that ranks. Blog posts, comparisons, tutorials, documentation that actually helps people.
Showing up where your customers are. Discord servers, subreddits, Twitter/X threads, niche forums, YouTube comments.
Building an audience before you launch. Give value for free for months. Then launch to people who already trust you.
Running cheap experiments. $50 on Meta ads. $50 on Reddit ads. See what messaging works. Iterate.
SEO from day one. Every page on your site answers a question someone is typing into Google. Target long-tail keywords with buying intent — "best AI writing tool for small business" not "what is AI."

Skip marketing and nothing else matters. The best AI-generated codebase in the world makes $0 if nobody knows it exists. I've watched brilliant builders ship incredible profitable AI apps and make nothing because they refused to learn marketing. I've watched mediocre builders print money because they understood distribution.

Pillar Two: Vibe Coding (Yes, You Need It — But It's Secondary for Making Money)

Let me be clear: you absolutely do need to be able to ship working software. You can't market vaporware forever. The vibe coding business model only works if there's an actual product behind it.

But here's the framework I want you to hold: if you ONLY learn coding, you'll become a vibe coder employee. You'll make 2x median salary. That's good money. You'll be comfortable. You won't be wealthy, and you won't have freedom. You'll still trade time for money — you'll just trade it at a higher rate.

The people making real money from AI coding tools aren't selling their coding hours. They're selling products that scale beyond their personal time. Product means: build once, sell many times. Service means: build every time, sell once.

What vibe coding gives you:

Speed. Ship a functional MVP in a weekend instead of a month.
Iteration velocity. Customer says "I wish it had dark mode"? Ship it in 20 minutes.
Reduced dependency. You don't need to hire a frontend person, a backend person, a DevOps person. You and an AI agent can do all of it.
The ability to run 4-10 product experiments per year instead of one.

The AI agents are genuinely incredible now. I use Hermes for complex orchestration — it's like having a senior engineer who works 24/7 and never complains. Cursor for in-editor coding. Claude for planning and architecture. The combination of neural net (your brain giving the commands) plus agent (the tool executing them) is unlimited. You think of the idea, describe it clearly, and the agent ships it. Then you iterate.

But here's what vibe coding won't give you: customers. Ideas validated by the market. Distribution channels. A brand. Recurring revenue from happy users. Those all come from the other three pillars.

Pillar Three: Vibe Support (Retention Is Everything)

Here's a stat that should scare you: it costs 5-7x more to acquire a new customer than to keep an existing one.

If you ignore support, you're lighting money on fire. Every customer who churns because they couldn't figure something out, or because a bug went unfixed for two weeks, or because they sent an email and nobody replied — that's a customer acquisition cost you'll never recover.

In the first few months, support is only you. No team. No help desk. No outsourced Philippines VA (yet). Just you, answering emails at weird hours, fixing bugs on Saturday night because a customer in a different timezone is blocked.

What good support looks like in a vibe business:

Reply within hours, not days. In the early days, aim for under one hour during waking hours.
Actually fix things. If a customer reports a bug, ship the fix. Don't "add it to the backlog."
Over-communicate. "Hey, saw your email — I'm looking at this now. Here's what I think is happening. I'll have a fix in the next 2 hours."
Build in public. Customer says "I wish this did X"? Tell them you're building X. Then actually build it. Then tell them you built it. That customer is now a evangelist.
Document everything. FAQ pages, help docs, onboarding videos. Every support question you answer manually is support debt. Pay it down by making the answer self-serve.

Support also gives you something money can't buy: product intuition. When you're the one reading every complaint, every feature request, every confused email, you develop an almost supernatural sense for what your product needs next. You'll know before any analytics dashboard tells you. That's gold. Don't outsource it too early.

Pillar Four: Vibe Bookkeeping (The Tax Man Is Real)

I know. This is the boring one. Nobody tweets about their chart of accounts. Nobody goes viral for properly filing quarterly estimated taxes.

But the tax man exists in every country, and he doesn't care that you're a solo vibe coder working from a coffee shop. You need to:

Accept payments. Stripe, Paddle, Lemon Squeezy, Gumroad. Pick one and set it up properly. Paddle and Lemon Squeezy handle VAT/sales tax for you globally, which is worth their higher fees when you're small.
Track everything. Revenue, expenses, refunds, chargebacks. Use something simple like Wave (free) or Xero. Don't use a spreadsheet — you'll mess it up and your accountant will hate you.
File taxes. Depending on your country, this might be quarterly estimated payments (US), VAT returns (EU/UK), or annual filings. Learn what applies to you. Set calendar reminders. Pay on time.
Separate accounts. Business bank account, business credit card. Never mix personal and business money. It makes taxes a nightmare and it makes you look unprofessional.
Plan for taxes. If you're in the US, mentally set aside 25-30% of every dollar that comes in. If you're in Europe with VAT, it's more complex — talk to an accountant.

I've seen people make $30k in their first year of a no-code AI business and then get destroyed by a tax bill they didn't plan for. Don't be that person. Bookkeeping is the least glamorous pillar, but it's the one that keeps you out of trouble.

The Four Pillars Work Together — Or They Don't Work at All

Here's how I think about it now:

Marketing brings people in the door.
Coding builds what they pay for.
Support keeps them paying.
Bookkeeping lets you keep the money legally.

If any one of these fails, the whole thing collapses. Amazing code with terrible marketing = no customers. Brilliant marketing with a broken product = refunds and bad reviews. Great product and marketing with no support = churn. All three working but no bookkeeping = the tax man takes it all.

People think vibe coding is just about coding. It's not. It's about building a business where AI handles the execution layer, and you handle the thinking, the distribution, and the customer relationships. That's the actual vibe coding business model — not "AI writes all my code and money appears."

What Is "Real Money" Anyway?

I want to define this because the internet has broken everyone's sense of what good money looks like.

Real money — "good money," the kind that changes your life — is about 3-4 times the median monthly salary in your region. Not San Francisco tech salary. Your actual region. If the median salary where you live is $2,000/month, "good money" is $6,000-$8,000/month.

Why that number? Because at that level, you can open the world. You feel like a human with options. You can say no to bad clients. You can take a month off to build your next thing. You're not rich — you're not buying a yacht — but you're free. And freedom is the actual goal. Not Lamborghinis. Freedom.

The math scales beautifully, by the way. 7,000 customers paying $1/month is $7,000 MRR. Seven customers paying $1,000/month is $7,000 MRR. The math is the same. The product, the marketing, the support — different. But the destination is identical. So stop obsessing over "should I build a $10/mo consumer app or a $500/mo B2B tool?" Build either. Just pick one and go. The formulas converge at the same revenue number — the only difference is how many customers you need to talk to.

This also means you don't need to be "the next Notion" or "the next Calendly." You need to find 50-500 people with a specific problem and solve it. That's it. That's the whole game. And AI makes finding and solving those problems faster than ever before.

How to Learn Without Burning Cash

Here's something I wish someone had told me four years ago: spend time, not money.

Time converts to knowledge at the best rate. Money converts to knowledge at a terrible rate. Every course, every bootcamp, every "AI entrepreneur accelerator" — they're selling you the fantasy I mentioned at the top. The actual knowledge comes from doing the thing.

AI tools are literally designed for teaching. You can ask Claude or ChatGPT "explain to me like I'm five what an API is" and get a perfect explanation. Then ask "now show me how to build one in Python." Then ask "why did you write it that way and not this other way?" You can have a conversation that teaches you more in two hours than a $2,000 Udemy course.

The key: don't go deep on technical internals. Every six months, everything resets. Six months ago everyone was obsessing over RAG architectures. Now AI agents handle that automatically. Studying "convolutional neural networks vs. transformers" when that knowledge will be fully automated in six months — that's not learning, that's procrastinating. It feels productive but it's not. Surface-level technical knowledge that lets you ship products is what matters. Deep technical knowledge that will be obsolete before you can use it is a waste of your time.

Here's the mental model I use: your neural net (your brain) gives commands. Your agent (Hermes, Cursor, Claude Code) executes them. Together they can build anything. The neural net doesn't need to know how the agent implements the solution — it just needs to know what good looks like, what the customer needs, and whether the output is correct. This combination is genuinely unlimited. A single person with this setup can out-ship a small dev team from 2019. I've seen it. I've done it.

How Long Does It Really Take?

Let me kill the "SaaS in a weekend" myth right now.

You need about six active hours per day minimum. That's 120+ hours per month. It's basically a university course load. If you have a full-time job, you're doing this on nights and weekends. That's fine — it just takes longer.

You'll need somewhere between four and ten product attempts before one works. Nobody knows which one. Nobody knows ahead of time. You might ship three absolute duds and then the fourth one — the one you almost didn't build, the one you thought was too simple — takes off. The only way to find out which one works is to ship all of them.

Free up five years in your head. Not because every attempt takes a year — most will take 1-3 months. But because the pattern recognition you need only comes from reps. Your first launch will be awkward. Your second will be smoother. By your fifth, you'll have systems. By your eighth, you'll have instincts. That takes calendar time, not just effort time.

Marketing alone needs a minimum of one month of focused effort per product to see real results. You can't launch on Monday and declare it dead on Friday. You need to push content, engage communities, run ads, iterate on messaging — for weeks. If you're not willing to give each product at least a month of marketing, don't start.

And here's the part that stings: learning doesn't stick the first time. You'll read about customer discovery, nod along, try it, and fail. Then six months later you'll re-read the same concept and it'll click differently because now you have context. That's normal. That's how learning actually works. It's not linear.

The Part Nobody Talks About

Alright, let's get uncomfortably honest.

Marketing always comes first. Before code. Before product. Before anything. I'll say it again because it's the thing most vibe coders skip and then wonder why they have zero customers. You can build the most elegant AI wrapper in existence. If nobody sees it, you have nothing. Start your marketing the day you start your product. Earlier, actually — start building an audience before you even know what you're building. Share your journey. Share what you're learning. By the time you launch, you should have people who want to see you win. That's marketing.

Your first three to four products will fail. That's the math. It's not pessimism — it's probability. Nobody's first product makes money. Nobody's. The people you see posting their success stories? You're seeing product number seven. They don't tell you about the first six. Every failure teaches you something specific: your pricing was wrong, your audience was wrong, your messaging was wrong, your timing was wrong. You need those lessons. You can't skip them.

Support is not optional. You built it. Only you understand how it works. Only you can support it for the first months. If you're not willing to answer customer emails at 11 PM for three months, don't build a SaaS. Build something that doesn't require ongoing customer relationships. The startup mythology about "move fast and automate everything" doesn't apply to customer support in the first year. Your early customers are doing you a massive favor by paying you and giving you feedback. Treat them like gold.

The tax man is coming. Accept it. Learn it. Don't ignore it. Every country has tax authorities and they are universally not fun to deal with. But they're predictable. The rules exist. Learn the rules and follow them. If you ignore taxes for two years and then try to fix it, you're looking at penalties, interest, and a world of stress. Just set up the systems from day one.

Can You Actually Make Money?

Yes.

But I want you to understand what you're actually signing up for. This isn't "learn to code with AI and get rich." This is: change the conditions of your life so that money can flow through them differently.

When I was a marketing director, I traded hours for salary. One source of income, directly tied to my time. The vibe coding business model breaks that. Now I have products that make money while I sleep. I have customers in countries I've never visited. I can work from anywhere because my "office" is a laptop and the AI agents I've trained.

It's not about passive income. Nothing is truly passive — you're always marketing, improving, supporting. But it is about decoupling your time from your income. That's the actual unlock.

Here's what I want you to take away:

The goal isn't to become a vibe coder. The goal is to become someone who can launch products repeatedly. Product launches are a skill. Like public speaking or sales. You get better at them. Each launch teaches you more about what works. After ten launches, you'll have pattern recognition that no course can teach you.

And launching products repeatedly requires all four pillars. You can't launch if you can't market (nobody shows up). You can't launch if you can't code (nothing exists). You can't launch if you can't support (everyone refunds). You can't launch if you can't do bookkeeping (you get fined into oblivion).

Start with marketing. Then coding. Then support. Then bookkeeping. In that order.

Build your first thing. Watch it fail. Learn why. Build the second thing. It'll fail differently. Learn why. By the fourth or fifth, something will click — not because you got lucky, but because you stopped making the mistakes you couldn't see before.

The AI coding tools exist. The agents exist. The payment processors exist. The distribution channels exist. Everything you need to make money vibe coding exists right now. If you want to understand how to monetize AI in practice, it's not about finding some secret trick — it's about combining these four pillars and executing consistently over years. The only missing piece is you deciding to start — and staying in the game long enough to get good.

FAQ

Do I need to know how to code before I start vibe coding?

No, but you need to be willing to learn. AI tools lower the barrier dramatically — you can ship working software with zero traditional coding knowledge. But you still need to understand logic, data flow, and how web apps work at a high level. The AI handles syntax and implementation; you handle architecture and product thinking. Expect to spend your first 2-3 months just learning how to communicate with the AI effectively. Prompting is a skill.

Which AI coding tools should I use for building a profitable AI app?

Cursor for in-editor coding (fast, integrated, great autocomplete). Claude or ChatGPT for planning, debugging, and architecture discussions. Hermes or similar agent frameworks for multi-step orchestration — tasks where you need the AI to plan, execute, and verify across multiple files or services. Start with Cursor + one LLM chat. Add agents when your projects get complex enough to need them.

How do I pick which product to build for my vibe coding SaaS?

Talk to people. Find a group you understand — maybe you worked in healthcare, logistics, education, whatever — and listen to what they complain about. Look for problems that people currently solve with spreadsheets, manual data entry, or "we have a guy who does this." Those are your targets. Don't try to build the next Facebook. Build something that saves a specific person 5 hours a week. Charge them for it.

How much money do I need to start?

Almost nothing. Domain name ($10-15/year). Hosting (Vercel/Railway free tiers, or $20/month when you scale). AI tool subscriptions (Cursor $20/month, Claude $20/month). Payment processor fees come out of revenue. You can be fully operational for under $60/month. The real investment is your time, not your wallet. Anyone trying to sell you a $5,000 "AI entrepreneur program" is selling you the shovel, not the gold.

What about AI wrapper income — is there still room?

Yes, massive room. The "AI wrapper" insult comes from people who don't understand business. Every SaaS is a wrapper — Salesforce wraps databases, Stripe wraps payment APIs, every mobile app wraps device hardware. The value isn't in the "wrapping," it's in the problem you solve and the experience you create. If you take an AI model and make it solve a specific workflow for a specific type of person — that's a business. Ignore the critics. Build something useful.

Why Nobody Is Buying: What Solo Founders Actually Need to Do

Axel Freeman — Thu, 30 Jul 2026 15:07:34 +0000

Why Nobody Is Buying: What Solo Founders Actually Need to Do

You built a product. Wrote the code, launched it, sitting there waiting. Silence. Zero sales. Sound familiar?

I've been through this. And now I'll tell you what I usually do when I launch products and when these kinds of questions come up. Let's go.

Building a product doesn't bring us closer to making money

The harsh truth is that creating a product doesn't get us closer to earning. What gets us closer: traffic, sales, landing pages, nice pages, videos, and more. Notice — to create all of this, we don't even need the product.

Unfortunately, building a product doesn't somehow bring us as businessmen any closer to making money. What brings us closer: traffic, sales, landing pages, nice-looking pages, videos, and so on. And notice — we don't even need a product to create any of that.

But nevertheless, people build products. And maybe that's right — it has value. But now we need to figure out where to get traffic for the first users, what to write, and what to do.

Let's take a step back to understand the general state of the playing field. Because hammering away at the question "where to get clients" — well, it's stupid. You don't understand at all. A beginner solo founder — he might not understand where clients are generated from.

We need to turn this from a black hole into something more or less coherent.

A mantra you need to remember

We are here for the money. We need to exchange our time for effort that our users will eventually value — and start paying us money for. That's the mantra. Always move toward the money.

Who the client is from the other side

On the other side of the product, our users are most likely completely random people. You should call them an audience. The audience needs to be counted.

Audience = people who made their first payment this month in the market you're participating in.

You need to understand this number. These are the people — the ones you'll be selling to.

Obviously, the number of people who will buy in six months is many times larger than those who will buy this month. But we're a startup — we want to test fast. So we focus only on people who will buy this month. That's the first number you need to calculate.

We've figured it out — most likely the number will be either very small or very large. Both are good. Based on this number, you can start picking free marketing tools.

If you have 300,000 clients per month — you have a broad market. The task: position yourself in a broad point of contact where clients get information about the product.
If you have a thousand clients per month — free outreach is your most important tool.

The maximum possible volume of clients should grow because new people reach the age where your product is needed. But that's background information. What matters is those who are buying this month.

Cosplaying the client: looking at the market through their eyes

When you cosplay the client — you start seeing through the eyes of the person who should buy your product. And that's when the picture of how this market is structured starts to emerge.

If the audience is broad (millions)

Look for entry points in social media. Instagram and so on. Look for bloggers who advertise similar products to a broad audience. Look through competitor hashtags — what posts, what they write. Maybe everything will become clear.

If it's a mass market

Search engines — Google is your friend. Google gives you the average slice of any domain. In the search results you'll encounter your potential sales channels: contextual advertising if your financial model allows it, rankings, third-opinion sites.

Third-opinion sites — rankings, reviews, alternatives. This is very important. The site needs to say that you're not bullshit. There will be some influencers, links to them. Looking through these platforms does this magical thing — it shows where transactions happen in the market.

If the product is complex and expensive

These guys don't need to attract traffic. Most likely they just need to search for contacts in the market. Forget about the product and just go talk to people. Use AI to find everyone's contacts and offer them a meeting. Say that you're an expert in this market, you know everything, and you're so fucking good that they have to reckon with you.

This might look naive and stupid. But if you write to the entire market with an offer to connect and nobody wants to talk — that's a red flag. It means there's no room for newcomers in this market. Because in a complex, large, competitive market, everyone is forced to adapt and react to incoming messages — nobody wants to lose everything.

Red flags

Can't find anything on Instagram about your audience — your product doesn't exist for clients right now. Just think about that. The situation: nobody is searching for this.
Can't find anything on Google — same thing.
Wrote to the entire market, nobody replied — the market is closed, no one cares.

The marketer's mindset

People are afraid to write

Afraid to write to strangers, afraid to look bad, or simply never done it, or expect incredible consequences from a single message. This gets in the way massively. By this, you're forbidding yourself from launching startups and attracting users.

The universal introductory mindset:

We have the right to take people's time through marketing. We have every right to spend a few minutes, hours, maybe even days of another person's time on viewing our marketing materials. The goal is one thing — convert.

Get rid of the fucking thoughts about "what will people think." Just think about it later. After 500 sends, ask yourself the question "what did people think." The answer will be very simple: no one cares. Most likely nobody even noticed you. And those 10 contacts who did notice — got interested, a dialogue happened, and some arrangements were made.

Forget all your life principles. Accept that we need the logic of a virus — one that doesn't ask questions, just copies and replicates further. I don't mean that we go through contacts like a machine sending the same message 100 times to 100 different people. That's genuinely what needs to be done. And here, the person gets in their own way first and foremost.

Put on the mask of a machine. The machine knows that today it needs to make 100 sends to relevant users. The machine will do it. That's it.

How to overcome yourself and write the first message

Take a pseudonym. Under a pseudonym, everything gets done a thousand times faster than under real personal data if you're truly scared.

Name yourself by a role: "CEO of the company," "sales manager." It doesn't matter that there's no name. Everything else you're broadcasting — that's the truth.
Invent a pseudonym — first and last name. You should have many names. Get a special name so it's not awkward. You're afraid for your own name, but if you use someone else's — you won't be afraid.
If you're Asian — take an English name. If you're English — take a Mexican or Spanish name. Translate your name into another language. That's fun, right?

Marketing is math. You invest yourself in the product. You invest yourself in the content. But not in delivery. You don't invest in delivery. All these emotions — irrelevant.

The magic number: 50 communications per day

In marketing, there is a number. 50.

50 communications per day. That's a thousand per month with 21 working days. On a regular startup schedule of 5/2 — consistently.

What counts as 50: 15 emails, 15 Telegram messages, 20 dictated posts or calls.

50 is the average norm, not the maximum level. You just need to learn how to build it into your life.

Why hasn't anyone come to me? Most likely because you're not doing anything for anyone to come.

Blindness

As humans, we have our own unique set of beliefs, speech patterns, vocabulary. The problem is that you sell to yourself perfectly — because you know yourself. But we can't know our audience thoroughly. There are no technologies yet to copy another person's consciousness.

So we will be blind. Segments at the start are a complete unknown, pitch-black chaos. You have to solve a problem that is clear to everyone: to grandma, to a child, to any person.

Blindness is needed so you don't play roulette. When you think you've come up with a unique style or a unique idea — you're playing roulette with a 99% chance of losing. That's how the world works. One old acquaintance will appreciate it, the rest won't give a shit.

It's important to be understandable. Don't use complex speech constructions. Don't think that your product only sells to right-handed people or only to business owners. You need universality, simplicity, and clarity. Basic vocabulary.

Blindness is needed so you don't hand out judgments. The only honest and professional way to evaluate marketing is to look at reports. How much traffic came, how many clients, how many actions were taken. There is no blindness here.

In style, speech, phrasing — blindness exists. This means accepting that your hypotheses are just little windows. You will need data. Being data-oriented means: remove bias, remove personal judgment. This is a tool that requires regular sharpening. You need to start right now — every month it will become more conscious.

Content: cheap, slow, works

The cheapest way at the start to get clients — and at the same time the most expensive in time — is to put out content.

Note: AI translates and stamps out text. The whole world is rolling toward everyone wanting to put out AI content. But it's sludge that AI itself doesn't need for indexing in the first place.

Use new ways of putting out content. For example, voice messages → text. I'm writing this message through voice messages right now. With simple tools, set yourself up so you put out content on all platforms where your audience is: Reddit — okay Reddit, Quora — okay Quora, something else — okay.

Record voice messages and put out content. Minimum one month, minimum 60 pieces. Wait for results for at least three months. Such a portion of unique content recorded personally by you — this is living flesh that all search engines chase after. This content needs to sit and eventually hit somewhere.

If you put out 90 pieces of content in a month on platforms that give free traffic and got nothing — that's another red flag. Maybe it's worth thinking not about how you're doing something wrong or don't know how to promote, but about the fact that the audience simply doesn't exist. Or it's inactive — everyone already bought long ago or has no interest, you can't convert them right now.

Blindness and evaluating results

When you have no users — you need at least one. Look at intermediate actions.

Following the 50 communications per day rule: 250 communications per week.

250 communications and 0 responses — the product isn't solving a real problem.
250 communications and there are responses — that's the vein of the business you need to explore. That's the pulse.

You post content and get reactions, likes, comments: you've stumbled onto a connection with the audience. It voted with attention.

You put out content — few views, incredible effort, nothing is happening: this is also a sign. Wrong path. Don't waste time on it.

You work with influencers — rejection everywhere, can't break through: maybe in this market, nobody needs them.

You pay money — people are enthusiastically ready to advertise: there's something here, the chance of success is higher.

You don't have to immediately look at money. Look at weak signals that say there's interest, views, reactions.

The second mountain

We never know how much time and effort it will take to attract the first paying user. This user will most likely cost hundreds of hours of marketing.

At the start, it seemed like 100% of the time should go into building the product, and after that traffic would appear. But behind the conquered peak of "build the product," a second peak appears — one not yet climbed. That's marketing.

Spend at least as much time on marketing as you spent on development. Then you can evaluate something.

FAQ

I built a product. Where do I start?
Find where your clients hang out. Look on Instagram, Google, forums. Didn't find them — the market might not exist. Found them — start publishing content. Minimum 2 pieces per day, minimum 3 months.

How long to wait for first sales?
Nobody knows. Rule: spend as many hours on marketing as you spent on development. After that, you can evaluate.

Sent 100 messages — 0 responses. What do I do?
Send another 150. 250 is the minimum sample. 250 and 0 responses — the product doesn't solve a problem. 250 and 5-10 responses — dig there.

Hire a marketer or do it myself?
At the start, only yourself. You know the product best. A hired marketer hasn't lived this pain. Hiring — only when you already have a working system and first sales.

How to tell if the problem is the product and not the marketing?
500+ touches and 0 signals. Bloggers refuse to even try it for free. Users who tried don't come back. All three signs — the problem is the product.

User Personas Are Dead: How to Actually Know Your Customer (OSINT Edition)

Axel Freeman — Mon, 27 Jul 2026 14:43:11 +0000

Originally published at axelfreeman.com

Buyer personas are fiction. Customer avatars are wishful thinking. ICP templates are corporate fanfiction. They all ask the same question — "who is my customer?" — and answer it with imagination instead of data. This article is about what comes after the Figma template. How to build a real user portrait: from Yandex Metrika to total session tracking to OSINT-level profiling that turns one email into 200+ data points. And where the legal line is.

All Roads Lead to One Question

User persona. Buyer persona. Customer avatar. Ideal Customer Profile. Jobs to Be Done. These are different words for the exact same thing: who is my customer, what do they look like, and what do they do?

When you know who comes to your product, you can start building relationships. Patterns emerge. Segments become visible. Every piece of data about your user is raw material for product ideas. And unlike your intuition, data doesn't lie. Mathematics beats subjective judgment every time.

The Rule: There is no useless information in understanding your user. The goal isn't a pretty persona card. The goal is to know more about your user than they know about themselves.

Phase 1: The Basics (What Everyone Tells You)

Install Analytics. Today.

Yandex Metrika. Google Analytics. AppMetrica for mobile. Connect it as an MCP server to your AI agent. Then ask your agent real questions: "How many users came today? Where did they go?"

Analytics gives you the first, most brutal truth: is anyone even there?

Track Everything

Every click. Every scroll. Every page. Every rage-click. Every form abandonment. Total surveillance of the user journey.

When you can replay a user's session and feed it to an AI for analysis, you start seeing patterns that dashboards hide.

Critical Distinction: Drop-off point vs Loss-of-interest point. A drop-off is when someone who intended to pay hits friction — fix it. A loss-of-interest point is when a curiosity browser naturally disengages — they were never a customer. Put them in a subscription funnel instead.

Phase 2: Where the Guides Stop

Every tutorial ends here: "Set up analytics, track events, hire analysts, good luck." But you can go much deeper. The information is already out there — publicly available, legally accessible.

Phase 3: Freeman Framework — OSINT-Level User Profiling

OSINT — Open Source Intelligence. All data has already been leaked or published. It's accessible through Google, public registries, and open APIs. One email address can become 200+ data points.

From an Email:

Registered sites: Which platforms have this email? GitHub? Stack Overflow? LinkedIn?
Professional identity: GitHub profile → coding languages, projects, activity patterns.
Business affiliations: Company registries → where they work, role, company size.
Public photos: Email → name → public photo → computer vision analysis.

From an IP Address:

Network type: Corporate? Home? Mobile?
Working hours: 2 PM from corporate IP = business user.
Organization targeting: Multiple users from one company network? That company is evaluating your product.

The Depth Potential: Classic approach: "user visited at 2 PM and left." Freeman Framework: "user visited at 2 PM from a corporate network, works at a 50-person fintech, codes Rust on GitHub, last tweeted about payment APIs."

Why: Hours = Money

User hours = revenue. More time in app = more value extracted. Every data point helps AI agents make better retention decisions.

The Legal Line

This is dangerous if you're sloppy. GDPR, CCPA, local variants. Rules:

Only open sources. Public websites, public registries, public APIs. Never breach paywalls.
Marketing purposes only. Build profiles to serve better, not to stalk.
Evaluate risk per jurisdiction.
When in doubt — lawyer. Real one, not a blog post.

The Bottom Line

User personas — Figma templates, stock photos, imaginary "Sally, 34, loves yoga" — are comfort food. Real understanding comes from:

Analytics: Metrika, GA, session replays.
Tracking: Total surveillance of the journey.
Distinction: Drop-offs vs. loss-of-interest.
Enrichment: Email → OSINT → 200 data points.
AI feeding: Let machines find patterns humans miss.

More data = better decisions = more engagement = more money.

Want to profile your users? Telegram: @axelfreeman. One analysis, free.

Marketing for Solo Developers: How to Get Your First Users (Without a Marketing Team)

Axel Freeman — Mon, 27 Jul 2026 14:43:10 +0000

Originally published at axelfreeman.com

Code is worthless now. Everyone can code. Your neighbor codes. Your friends code. The real competition has moved to marketing — and "marketing" is a garbage word that means everything and nothing. Like "computer." This article is me untangling that mess. You'll learn why being clear beats being clever, how to legally steal competitor customers, and why $300 is all you need to validate any idea. No marketing team. No bullshit.

The Code Is Dead. Long Live Marketing.

Here's the uncomfortable truth: writing code used to be a superpower. It's not anymore. Every person in your network can build something now. AI makes it trivial. Your competitive advantage can't be "I built a product" — everyone has that.

The next level of competition is marketing. And I don't mean "marketing" the way people use it as a catch-all for "stuff that makes people buy things." That's the problem. "Marketing" is like the word "computer" — it's so broad it's become meaningless.

If you're a solo developer (or a tiny team) reading this, your goal is simple: get your first users. Not build a marketing empire. Not hire a CMO. Just find people who will pay for what you built.

The Goal: We're not building a marketing machine for a century. We're showing up to the market with determination and audacity, stealing a few customers from competitors, and seeing what happens. That's it.

Eyes and Ears: The Only Channels That Matter

Let's shrink the problem immediately. You're behind a screen. Your customer is behind a screen. What human senses can you actually reach?

Two. Just two. Sight and hearing. Eyes and ears.

Sight (eyes): Text, images, video, UI, design, ads
Hearing (ears): Audio, podcasts, voice, music, video soundtracks

Every platform ever built works through exactly these two channels. Text is the cheapest, simplest, and most universal format. Text becomes voiceover scripts. Scripts become video. Everything starts with text.

Don't Segment When You Have Zero Users

There's a million frameworks about audience segmentation. But when you have zero users, segmentation is poison. The source of your "persona" isn't reality — it's someone else's opinion recycled from the internet.

On day one, the only source of truth is market response. Traffic comes in. People interact. Some convert. Some don't. That data is real. Everything else is guessing.

The Segmentation Trap: When you narrow your message to "grandmas in Odessa," 90% of people who could have converted scroll past you. Write broad. Write clear. Let the market tell you who your customers are through actual data.

The Math of "Where": Count the People Buying Right Now

"Where do I find users?" starts with the word where. It's a location question. And the location is: wherever people are currently making the decision to buy something like your product.

Your first job: estimate how many people today made the decision to buy a product like yours — somewhere else. You don't need precision. You need a ballpark. Use AI for desk research.

These people — the ones making a buying decision right now — are your first base. They could have come to you. They didn't. Let's fix that.

Steal Your Competitors' Customers (Legally)

Here's the uncomfortable strategy that works: steal customers from competitors.

People are inattentive. They glance at search results. They misread brand names. Bid on competitor brand keywords in search ads. Show up when someone types your competitor's name into Google.

Is this legal? Yes. Millions of court cases across jurisdictions have established that bidding on competitor trademarks is fair game. Check your local regulations.

Why Stolen Customers Don't Lie: A customer who clicked your ad expecting a competitor already came to pay. If they don't convert — that's real data. Your product has a problem. Not your marketing.

Being Clear: The Only Marketing Skill That Actually Matters

First, accept this: people are selfish. Every second they think about themselves. Your marketing has to work with this.

Your message needs to answer exactly three questions, immediately:

What service do you provide?
What problem does it solve?
What does the person get by coming to you?

Vocabulary: use the basic dictionary. Cambridge Dictionary level. Agents can do this — just tell them "write at B1 vocabulary level."

Tone of Voice: Be Beige

Your tone of voice is not a creative decision. It's a research task. Scan 10-20 competitors. Ask an AI to analyze their average tone. That's your tone. It'll be beige. Boring. But native to the market.

Uniqueness Is a Trap

There's a fantasy that being "unique" and "disruptive" wins on launch. It doesn't. Uniqueness on day one means a 99% failure rate.

Be approximately average in presentation. Win on execution.

Buy Traffic: $300 Validates Any Idea

The simplest, fastest method: buy traffic.

Budget formula:

Ask an AI: "What's the average cost per lead in [your market]?"
Take the upper estimate. You're new. Competitors outbid you.
Multiply by 10. That's your validation budget.

This number is almost never more than $300. Any idea. Any market.

Do It Yourself. With AI Agents. For validation, AI agents handle it. Create ads. Write copy. Analyze results. All solo.

Competition Is Always There

You are not alone. Thousands of people worldwide are testing your market right now. There's no such thing as "no competition."

But being small means you can spy. Competitors' landing pages, ads, pricing — all public. Use AI to analyze them.

The Bottom Line

Marketing for a solo developer isn't about building a brand. It's about survival through clarity and audacity.

Checklist:

Estimate your market. How many people are buying today?
Be clear. Basic vocabulary. Three questions. Beige tone.
Don't segment. Zero users = zero data. Write broad.
Buy competitor traffic. Bid on their keywords. It's legal.
Spend $300. Validate.
Do it yourself. AI agents are your marketing team.

Most startups die from overthinking and underspending. You now know enough to do neither.

Need help? Telegram: @axelfreeman. One neural network search problem solved for free.