DEV Community: extebarrri

I Replaced 3 SaaS Subscriptions with Free AI Tools and Saved $852/Year

extebarrri — Sat, 21 Mar 2026 19:15:19 +0000

$86 per month. $1,032 per year. That's what I was paying for three SaaS subscriptions I barely used.

In January 2025, I cancelled all three and replaced them with free AI tools. Six weeks in, I haven't missed any of them.

The Problem: SaaS Creep

I had:

Formulas add-on ($12/month)
Descript ($24/month)
Beautiful.ai ($50/month)

I used each maybe 2-3 times per month. They auto-renewed and I forgot they existed.

Subscription 1: Formula Writer → Claude

What I was paying for: Type "sum the revenue column" and get a formula.

What I switched to: Claude 3.5 Sonnet + a prompt template.

Feature	Old	New
Speed	15 seconds	5 seconds
Accuracy	85%	98%
Languages	Excel only	Excel, Sheets, SQL, Python
Cost	$12/month	$0.03/month

Cost savings: $143.64/year

Subscription 2: Descript → Whisper + Claude

What I was paying for: Upload a video, auto-generate captions.

What I switched to: Whisper API + ffmpeg + Claude.

Feature	Old	New
Speed	2 minutes	1 minute
Accuracy	92%	96%
Formatting	Professional	Manual
Cost	$24/month	$0.01/month

Cost savings: $287.88/year

Subscription 3: Beautiful.ai → Claude + Google Slides

What I was paying for: Professional pitch templates + auto-formatting.

What I switched to: Claude for structure + Google Slides for design.

Feature	Old	New
Time to deck	20 minutes	40 minutes
Cost	$100/month	$0

Cost savings: $1,200/year

The Real Numbers

Before: $86/mo ($1,032/year)
After: $15/mo ($180/year)
Annual savings: $852

What I'd Still Pay For

Figma ($12/month): No free alternative
Linear ($10/month): Saves 5 hours/week
Vercel: Deployment worth paying for

Lessons Learned

Free AI tools are good enough for individual tasks. You need technical chops to DIY. Defaults matter. SaaS marketing works.

What SaaS are you overpaying for?

Free SEO Tools I Use Daily as a Solo Founder Running 15 Sites

extebarrri — Sat, 21 Mar 2026 19:08:46 +0000

I manage 15 websites as a solo founder. I've also got 47 blog articles across these sites and zero budget for Ahrefs or Semrush ($200-500/month).

So I built a SEO stack from free tools. No paid SaaS. No shortcuts. Just discipline.

The Essential Five

1. Google Search Console (GSC)

This is non-negotiable. It tells you:

Which pages Google can find
Why it can't find others
What people search for to reach you
How often you appear in results
Core Web Vitals issues
Security problems

My morning routine:

Check Coverage report
Check Performance report for new keywords
Check Security & Manual Actions

Takes 5 minutes. Catches everything important.

2. PageSpeed Insights + Web Vitals

Google's ranking algorithm cares about:

Largest Contentful Paint (LCP): < 2.5s
First Input Delay (FID): < 100ms
Cumulative Layout Shift (CLS): < 0.1

I run this before publishing every article. If I see red flags, I optimize before hitting publish.

3. CalcSEO for Traffic Estimation

Before I write a 2,000-word article, I want to know: is there enough search volume?

CalcSEO shows monthly search volume, ranking difficulty, and traffic opportunity.

4. SEO Meta Analyzer

Every page needs a unique title, meta description, H1 tag, and internal links.

I run this on every blog post before publishing. It's 2 minutes of checking.

5. ToolScout for Discovering New Tools

ToolScout is how I find other free tools instead of scrolling Reddit for hours.

The Numbers

Across 15 sites:

47 published articles
15,200 monthly organic impressions
1,120 monthly organic clicks
6.8% average CTR (industry average is 2-3%)

My content ranks for:

340+ unique keywords
22 keywords in top 3 positions
89 keywords in top 10 positions

Cost to achieve this: $0/month for tools.

What I Don't Use

Ahrefs ($99+/mo), Semrush ($99+/mo), Moz. Too expensive or outdated.

Conclusion

You don't need a $200/month SEO subscription. You need Google Search Console, PageSpeed Insights, CalcSEO, SEO Meta Analyzer, and discipline.

What's your SEO tool stack?

Building AI Agents in 2026: Templates, Evaluation, and Production Lessons

extebarrri — Sat, 21 Mar 2026 19:08:37 +0000

Two years ago, building an AI agent meant assembling your own orchestration layer, writing prompt templates by hand, and praying your tool calling worked. Today in 2026, it's a commodity.

I've shipped 12 production agents in the last 8 months. Here's what I learned about templates, evaluation, and avoiding expensive mistakes.

Build vs. Use a Template

The question isn't "Can I build an agent?" It's "Should I?"

I built from scratch for:

Custom domain logic
Unique tool integrations
Proprietary evaluation criteria

I used templates for:

Retrieval-augmented generation (RAG)
Customer support agents
Internal documentation assistants
Lead qualification chatbots

AgentKit saved me 16 hours on my fifth agent. It includes prompt templates, tool calling scaffolding, evaluation harness, and deployment configs.

Evaluating Agent Quality

I evaluate every agent across four dimensions:

1. Accuracy — Benchmark against gold-standard answers, measure semantic similarity
2. Latency — p50, p95, p99 response times
3. Cost — Cost per query (LLM API calls + token usage)
4. Safety — Does it refuse dangerous requests?

I run evaluation before every deployment. If accuracy drops below 90% or safety below 95%, I don't deploy.

Common Pitfalls

Pitfall 1: Using the wrong model
I tested Haiku for lead qualification. Accuracy was 73%. Switched to Sonnet—96% accuracy. The cost went up but paid for itself in one day.

Pitfall 2: Prompt injection
My support agent could be manipulated into revealing customer data. Solution: Separate user input from system context with explicit delimiters.

Pitfall 3: Hallucination at scale
An agent works great in testing, then produces garbage in production. Training data mismatch. Solution: sample 1% of production queries and manually verify accuracy weekly.

Pitfall 4: Runaway costs
I had an agent that retried failed tool calls infinitely. Added a max retry count and cost budget.

Template Comparison

Dimension	Build from Scratch	Use Template
Time to deploy	2-3 weeks	2-3 days
Accuracy out of box	60-75%	85-95%
Cost to maintain	High	Low
Best for	Custom domain logic	Standard use cases

Production Lessons

Test with real data
Monitor hallucination rate
Version your prompts
Budget for inference
A/B test models

AI agents are solved problems for common use cases. The decision isn't whether to build—it's whether to template or customize.

What's your evaluation framework?

How I Monitor MCP Servers in Production — Tools and Lessons Learned

extebarrri — Sat, 21 Mar 2026 19:03:26 +0000

Three months ago, one of my MCP servers crashed at 2 AM on a Friday. I didn't find out until Monday morning when a customer opened a support ticket. By then, 60+ API calls had failed silently, and I'd lost two days of data.

That's when I realized: MCP servers have no built-in observability. They fail quietly. There's no error dashboard, no alerts, no uptime tracking.

I spent 8 weeks building a monitoring stack for MCP servers. Here's what I learned.

The Problem

Unlike traditional SaaS APIs, MCP servers often:

Run on VPS with minimal logging
Don't have native error tracking
Fail gracefully but return garbage responses
Have no built-in health check endpoints
Don't expose metrics in standardized format

What to Monitor

After my 2 AM incident, I identified five critical metrics:

Uptime & Availability — Is the server actually handling requests?
Error Rates — What percentage fail? I set 2% threshold.
Response Times — p50, p95, p99 latency
Token Usage — MCP servers burn tokens fast
Resource Utilization — Memory, CPU, disk

Real Lessons Learned

Lesson 1: Server crashes aren't your biggest problem.
The real problem was the server with 500ms latency for 3 days. Customers quietly switched competitors.

Lesson 2: Logs are not metrics.
You can't alert on logs. Use structured metrics instead.

Lesson 3: Test your health checks.
My health check always returned green, even when the core service failed.

Lesson 4: Set realistic thresholds.
My 1% error threshold got me paged 40 times/day. Now 5%.

Conclusion

Monitoring isn't sexy. But it's the difference between customers discovering outages and you discovering them first.

Cost: 2 hours setup + $50-100/month infrastructure. Upside: Not woken at 2 AM.

What's your monitoring approach?

Test Article

extebarrri — Sat, 21 Mar 2026 19:02:12 +0000

Test

This is a test.