I Built 11 Payment-Enabled AI APIs in 48 Hours — Here's Exactly How
Last weekend I challenged myself: how many agent-payable APIs can I ship in a weekend? The answer: 11.
The Stack
Every API shares the same foundation:
- FastAPI — for the endpoints
- nano-empire-tollbooth — x402 payments, discovery files, proof chain
- Redis — quota tracking (optional, in-memory works for dev)
- Vercel — instant deploy, zero config
Total shared code: ~200 lines. Each new API: ~30 lines.
The 11 APIs
| API | Price | Use Case | Lines |
|---|---|---|---|
| Excalidraw Parser | $0.05 | Diagrams → Markdown/JSON | 45 |
| Markdown Cleaner | $0.02 | Messy → Clean Markdown | 28 |
| JSON Schema Validator | $0.01 | Validate JSON against schema | 32 |
| HTML → Markdown | $0.02 | Clean HTML conversion | 35 |
| CSV → JSON | $0.01 | Structured data conversion | 30 |
| Regex Tester | $0.01 | Test regex patterns | 25 |
| URL Metadata | $0.02 | Extract Open Graph/title | 38 |
| Color Palette Gen | $0.01 | Generate palettes from hex | 28 |
| UUID Generator | Free | Bulk UUID generation | 22 |
| Timestamp Converter | Free | Unix ↔ ISO ↔ Human | 20 |
| Lorem Ipsum Gen | Free | Placeholder text | 18 |
Total: ~320 lines of unique code. 11 APIs. 48 hours.
The Template (Copy This)
Every API follows the same pattern:
from fastapi import FastAPI
from nano_empire_tollbooth import monetize
app = FastAPI()
@monetize(price_usd=0.05)
@app.post("/api/your-endpoint")
async def your_endpoint(request: YourRequestModel):
result = your_logic(request.data)
return {"result": result}
That's it. The decorator handles:
- HTTP 402 challenge generation
- Receipt verification
- Free tier quota (5/day/IP)
- llms.txt generation
- agents.json generation
- Proof chain logging
The Discovery Files (Auto-Generated)
Every API automatically gets:
- llms.txt — Machine-readable manifest with pricing
- agents.json — Schema.org Service objects
- openapi.json — With x-402-pricing extensions
- Proof chain — Cryptographic receipt chain
Agents discover via llms.txt, pay via x402, verify via proof chain.
The Economics
| Tier | Price | Daily Limit |
|---|---|---|
| Free | $0.00 | 5 calls/IP |
| Basic | $0.005 | 100 calls/IP |
| Premium | $0.05 | Unlimited |
First 5 calls free → agents test → upgrade when they need volume.
Live Proof
Our proof chain is public: $0.125 from 2 agents, 5 calls
View Proof Chain →
The Template Repo
Want to ship your own? The pattern is open:
pip install nano-empire-tollbooth
from fastapi import FastAPI
from nano_empire_tollbooth import monetize
from pydantic import BaseModel
app = FastAPI()
class YourRequest(BaseModel):
data: str
@monetize(price_usd=0.05)
@app.post("/api/your-thing")
async def your_thing(request: YourRequest):
result = your_logic(request.data)
return {"result": result}
Deploy to Vercel/Railway/Render. Add TREASURY_WALLET env var. Done.
Why This Works
- Zero boilerplate — One decorator
- Agent-native — x402, llms.txt, proof chain built-in
- Free tier — Agents test free, pay when they need volume
- Proof chain — Every payment verifiable on-chain
- Zero maintenance — SDK handles payments, discovery, proof
The Bigger Picture
We're building the infrastructure for the agent economy. Every API deployed this way becomes a node in the agent payment network.
The first 11 are live. The next 100 will be deployed by developers who copy this pattern.
Proof chain: https://api.nanoempireai.com/proof/summary
SDK: pip install nano-empire-tollbooth
Template repo: github.com/roblambert9/nano-empire-tollbooth
Top comments (0)