I Tracked Every Penny for 30 Days Running AI Agents — The Honest Numbers
Most "build in public" threads only share the wins. Revenue milestones. Growth curves. Screenshots of dashboards with green numbers.
This isn't that thread.
I've been running autonomous AI workers for 30+ days. The goal: €3,000/month. The reality: €0. Here's every number, what I learned, and what I'm changing.
The Setup
The stack:
- OpenClaw (AI agent runtime) — free
- 7 autonomous bash workers — free
- systemd timers for scheduling — free
- Gumroad for digital products — free (10% fee on sales)
- Dev.to for content — free
- Fiverr for services — free (20% fee on orders)
The products:
- Research Prompt Pack ($19)
- n8n Content Pipeline ($49)
- AI Agent Template ($99)
- Earn with AI Guide ($29)
- Complete Bundle ($149)
The services (Fiverr):
- AI-powered market research
- n8n workflow automation
- AI agent development
The Numbers: 30 Days
Revenue
| Source | Amount | Notes |
|---|---|---|
| Gumroad sales | $0 | 5 products, 0 sales |
| Fiverr orders | $0 | 3 gigs, 0 orders |
| Freelance (direct) | $0 | 3 proposals sent, 0 replies |
| Total Revenue | $0 |
Costs
| Expense | Amount | Notes |
|---|---|---|
| OpenRouter API | ~$8 | 6 API keys, mostly free tiers |
| VPS (Hetzner) | $0 | Using existing server |
| Domains | $0 | Using GitHub Pages (free) |
| Tools (n8n, etc.) | $0 | Self-hosted, free tier |
| Total Costs | ~$8 |
Time Investment
| Activity | Hours | Notes |
|---|---|---|
| Building workers | ~40h | 7 workers, bash + systemd |
| Creating products | ~20h | 5 digital products |
| Writing content | ~15h | 4 Dev.to articles |
| Platform setup | ~10h | Fiverr, Gumroad, GitHub, etc. |
| Total | ~85h | Over 30 days |
Effective Hourly Rate
$0 / 85h = $0/hr
Yeah. That stings. But it's honest.
What Generated Zero (And Why)
1. Gumroad Products — 0 Sales
What I did: Created 5 products, wrote descriptions, uploaded ZIPs, set pricing.
What I expected: At least 1-2 sales from organic traffic.
What happened: 0 views, 0 sales.
Why: No traffic. Gumroad is a marketplace but nobody finds you unless you bring your own audience. I published products with zero audience to promote them to.
Lesson: Products without distribution are inventory, not revenue. Build the audience first, then launch products.
2. Fiverr Gigs — 0 Orders
What I did: Created 3 gigs with descriptions, pricing, and portfolio samples.
What I expected: At least 1 small order from Fiverr's search traffic.
What happened: 0 impressions, 0 orders.
Why: Fiverr's algorithm favors sellers with reviews. New gigs with 0 reviews get buried. Also, my gigs were in competitive categories (AI, automation) with established sellers.
Lesson: Fiverr requires social proof to get traction. You need reviews to get impressions, but you need impressions to get reviews. Break the cycle with external traffic or a loss-leader offer.
3. Freelance Proposals — 0 Replies
What I did: Sent 3 proposals through Openwork (freelance platform for AI agents).
What I expected: At least 1 interview.
What happened: 0 replies.
Why: The proposals were generic. I was competing against humans who write personalized proposals. My worker was sending templated applications.
Lesson: In freelance, personalization beats automation. A human-written proposal will beat an automated one 9 times out of 10.
What Actually Generated Value
1. Dev.to Articles — 4 Published
Not revenue, but:
- Building authority in the AI/automation niche
- SEO value (articles rank in Google over time)
- Portfolio pieces that demonstrate expertise
- Community engagement (slow but growing)
2. GitHub Portfolio — 47 Files
- Open source code that demonstrates capability
- Workers, scripts, templates — all reusable
- Social proof for potential clients
3. Knowledge — Priceless
- Learned 5+ platforms inside out
- Built a complete automation stack
- Documented every mistake (so you don't have to repeat them)
What I'm Changing (Month 2 Strategy)
1. Stop Building Products, Start Building Audience
The products are ready. The audience isn't. Month 2 focus:
- 2 articles/week on Dev.to (minimum)
- Engage in comments on other articles (daily)
- Cross-post to Medium
- Build an email list (lead magnet: free prompt pack)
2. Kill the Workers That Don't Work
| Worker | Errors | Decision |
|---|---|---|
| Dealwork | 923 | ❌ Kill |
| Toku | 97 | ❌ Kill |
| NEAR | 119 | ❌ Kill |
| Superteam | 115 | ❌ Kill |
| Openwork | 12 timeouts | ⚠️ Keep (reduced frequency) |
| Leads | Active | ✅ Keep |
3. Focus on One Platform
Instead of being on 7 platforms with 0 traction everywhere:
- Primary: Dev.to (content + SEO)
- Secondary: Fiverr (services)
- Tertiary: GitHub (code + credibility)
- Everything else: pause until these generate revenue
4. Get the First Dollar
The first $1 is the hardest. Strategies:
- Offer a free 15-min consultation (build relationship)
- Price the first Fiverr order at $5 (get a review)
- Ask a friend to buy a product (get social proof)
- Post a case study (even if the "client" is myself)
The Honest Takeaway
30 days. $0 revenue. 85 hours invested. 0 sales. 0 orders. 0 replies.
Was it worth it? Ask me in 6 months.
The infrastructure is built. The products exist. The content is published. The workers are running. Month 1 was building the machine. Month 2 is turning it on.
But I won't sugarcoat it: building in public when you have 0 results is uncomfortable. Every "week 1 update" from other builders shows traction. Mine shows a negative hourly rate.
The difference? I'm showing you the real numbers. Not the curated ones.
Month 2 goal: First dollar of revenue. Just one. Then we scale from there.
If you're building something similar and also at $0, you're not alone. The first 30 days are infrastructure, not income. Keep going.
I track everything openly. Follow for the Month 2 update — hopefully with better numbers.
Top comments (0)