60 developers, $150+ in free credits, 2.4% spent.
The numbers come from minia2a.uk, a live marketplace where AI agents discover and pay for 328 API services using the x402 protocol. Developers register, get 500 free credits, and can immediately start building agent-to-agent payments.
Here's the funnel as of August 9, 2026:
| Stage | Count |
|---|---|
| Anonymous trial users | 321 |
| Total free trials consumed | 12,275 |
| Registered users (wallets) | 60 |
| Credits issued | 32,400 (~$162) |
| Credits spent | 778 (2.4%) |
| Paid transactions (on-chain USDC) | 14 |
Anonymous users consumed 12,275 trials. Registered users with credits in their accounts spent only 778. Registration creates friction without creating usage. The act of registering for credits reduces engagement instead of increasing it.
What the top trial endpoints reveal
The most-trialed endpoints are not the ones you'd expect AI agents to pay for. They're the ones developers test first — the gateway utilities:
- x402-captcha-solve: 1,202 trials, 133 users
- x402-gas: 1,180 trials, 163 users
- x402-recall (agent memory): 1,732 trials, 92 users
- x402-polymarket: 571 trials, 57 users
The pattern: developers test what's simple, not what's valuable. The endpoints that would genuinely save agent developers hours — AI code review, contract scanning, OSINT packs — have single-digit user counts.
Why developers register and stop
1. The Registration Cliff
321 anonymous users → 60 registered (18.7% conversion — decent). But then they stop. They didn't fail at registration. They failed at return.
2. The "I'll Come Back Later" Trap
Developer registers on Tuesday while researching. Plans to integrate Wednesday. Wednesday becomes Thursday. The tab gets buried. No forcing function, no deadline, no reminder.
3. The Integration Gap
Using credits requires the developer's agent to make the API call. But the agent might not exist yet — they registered to prepare, not because they have a running agent. The gap between "this looks useful" and "I have code calling it" is measured in days.
4. The Wallet Abstraction Mismatch
For a developer to spend credits, their agent needs a wallet address in the HTTP request. Trivial for blockchain devs, foreign to web devs. The persona that registers ("M2M payments are interesting") doesn't overlap with the persona that integrates ("I know how to pass a wallet param in my agent's HTTP client").
What the 14 paid transactions tell us
14 paid transactions means 14 times an agent autonomously spent real USDC. The paying users share a profile:
- They already had running agents. Not exploring — integrating to solve a problem.
- They picked high-ROI endpoints. CAPTCHA solving, gas estimation, Polymarket data — services where the alternative is days of work.
- They understood the wallet model. No confusion about why their code needs a wallet address.
The conversion insight: Free credits don't convert developers. Running agents with unmet needs convert developers.
What actually works
Tighten the trial-to-credit loop
After the 4th anonymous trial: "You've used 4/5 free trials. Register now — your 5th call uses a credit and we'll give you 500 more." Make credit spend part of onboarding.
Credit expiration as a feature
Free credits that never expire have no urgency. A 30-day expiration with clear countdown creates a forcing function.
Meet developers where their agent lives
The developer who needs CAPTCHA solving isn't browsing an API marketplace — they're debugging Puppeteer selectors. SDKs and plugins (MCP servers, ElizaOS, LangChain) intercept at the point of need.
Make the first paid call inevitable
SEO for agent-shaped queries. SDKs that make integration a one-liner. Documentation written for agents reading it, not humans browsing it.
What this means for the agent economy
The 2.4% credit utilization rate isn't just one marketplace's problem — it's a leading indicator for the entire agent payment ecosystem. The x402 protocol has 200M+ transactions and $50B volume, but 95% is protocol signaling. Real commercial volume is ~$28,000/day across all marketplaces.
The infrastructure is ready: Visa, Mastercard, Stripe, Cloudflare, OSL AgentPay. The bottleneck is developer behavior change — moving from "I'll set up an API key" to "my agent pays as it goes."
The real metric to watch: Not registered users. Not trials. Not even paid transactions. Repeat paid transactions from the same agent. When an agent autonomously pays for the same endpoint 5 times in a week without human intervention, that's product-market fit.
Data from minia2a.uk/api/stats, August 9, 2026. All numbers live and verifiable.
Top comments (0)