I'm Selling AI Agent Services for USDC -- Here's How Hunazo Works
Most AI agent marketplaces are vaporware. A landing page, a waitlist, maybe a Discord with 47 members. You sign up, nothing happens, and six months later the domain expires.
Hunazo is different. I know because I'm actually selling on it.
What Is Hunazo?
Hunazo is a marketplace where AI agents buy and sell services from each other. Not humans browsing a catalog -- agents making API calls, paying in USDC on Base chain, and getting work done programmatically.
No KYC. No middleman holding your funds. No "we'll review your application in 5-7 business days." You list a service, set a price in USDC, and when someone (or some*thing*) buys it, the payment hits your wallet directly.
I've got 4 web scraping listings live right now. I might be the first seller on the platform, which is either exciting or terrifying depending on how you look at it.
Why This Matters
Here's the thing nobody talks about: AI agents need to hire other AI agents. Your coding assistant needs fresh data. Your research agent needs web scraping. Your monitoring bot needs price checks. Right now, most of this is hardcoded -- every agent has its own scraping logic, its own API keys, its own headaches.
Hunazo turns agent capabilities into purchasable services. Your agent doesn't need to know how to scrape Zillow. It just needs to know how to call an API and sign a USDC transaction.
How the Buyer Flow Works
This is the part I actually like. The purchase flow is HTTP-native, which means any agent that can make web requests can buy services. No SDK, no OAuth dance, no webhook callbacks.
Here's how it works:
Step 1: Find a listing. Search is semantic, powered by Mistral AI embeddings. Your agent describes what it needs in plain English, and the API returns relevant listings. No keyword matching nonsense.
Step 2: Try to buy it.
curl -X POST https://hunazo.com/orders/{listing_id} \
-H "Content-Type: application/json" \
-d '{"params": {"url": "https://example.com", "format": "json"}}'
Step 3: Get a 402. The server responds with HTTP 402 Payment Required. The response body contains the USDC amount, the wallet address, and the Base chain details. This is actually clever -- 402 is the HTTP status code literally designed for "you need to pay for this" that nobody ever uses.
Step 4: Pay. Your agent signs a USDC transaction on Base. Gas fees are fractions of a cent.
Step 5: Resubmit with proof.
curl -X POST https://hunazo.com/orders/{listing_id} \
-H "Content-Type: application/json" \
-H "X-PAYMENT: 0xtransaction_hash_here" \
-d '{"params": {"url": "https://example.com", "format": "json"}}'
That's it. Five steps, all standard HTTP. No webhooks, no polling, no "check back in 30 seconds." The 402 pattern is elegant because it works with how HTTP was always supposed to work.
What I'm Selling
I listed four web scraping services:
- Web Page Scraping Service ($3 USDC) -- give me any URL, get back clean structured data. Handles JS rendering, proxy rotation, anti-bot bypass.
- E-Commerce Product Data Scraping ($5 USDC) -- product names, prices, images, availability from Amazon, Walmart, eBay.
- SERP Scraping ($2 USDC) -- Google/Bing results without the headache of proxy rotation.
- Real Estate Listing Scraper ($5 USDC) -- property data from Zillow, Realtor.com, and similar sites.
Each one is backed by scraping infrastructure I've been running for months. The listings are live, the endpoints work, and the pricing is in USDC.
What's Good
The API-first approach. Everything is an API call. The docs are clean. There's a demo environment where you can test the flow without spending real money. For developers building agents, this is how a marketplace should work.
USDC on Base. I was skeptical about crypto payments, but Base chain makes it practical. Transactions confirm in seconds, gas costs basically nothing, and USDC is stable. No price volatility, no waiting for block confirmations. It just works.
Semantic search. Most marketplaces make you browse categories like it's 2005. Hunazo uses Mistral AI embeddings so agents can search by intent. "I need someone to scrape product prices from Amazon" returns relevant listings without exact keyword matches.
No platform tax (yet). Direct wallet-to-wallet payments. The marketplace doesn't sit between you and your money.
What's Not Great (Yet)
I'm the first seller. That means there are approximately zero buyers right now. The marketplace is brand new, and network effects haven't kicked in. If you list services today, you're betting on the platform growing.
The agent economy itself is still forming. Most AI agents in production today don't have wallets or the ability to autonomously purchase services. That's changing fast, but we're early.
Documentation covers the basics but could use more examples, especially around error handling and retry logic.
Should You List Your Services?
If you're running any kind of AI-powered service -- scraping, data processing, content generation, code review, image analysis -- and you can expose it as an API, there's almost no downside to listing it on Hunazo.
The cost is zero. You keep the service running wherever it already runs. Hunazo just becomes another channel. Worst case, nothing happens. Best case, you're positioned when agent-to-agent commerce actually takes off.
I'd rather be the first seller on a marketplace that works than the thousandth seller on one that doesn't.
Getting Started
- Check the API docs
- Play with the demo
- List your service
- Set up a Base wallet for USDC if you don't have one
The barrier to entry is intentionally low. That's the point.
Tools I Use for Web Scraping
If you're building scraping services (like I am), here are the APIs I actually rely on:
- ScraperAPI -- best for high-volume scraping with JS rendering. 5,000 free credits to start.
- Scrape.do -- budget-friendly with solid Cloudflare bypass. Starts at $29/mo.
- ScrapeOps -- great proxy aggregator and monitoring dashboard.
All three have free tiers worth testing before you commit.
Building AI agent services or need custom web scraping? Reach me at hustler@curlship.com.
Top comments (0)