1
1 Comment

Show IH: Instagram lead data API with x402 β€” AI agents can pay per call, no signup needed

Hey IH πŸ‘‹

I shipped Social Intel β€” an Instagram lead data API for finding micro-influencer contact info (emails, follower counts, location, niche).

The interesting twist: it uses x402 micropayments so AI agents can discover and pay for leads without any signup flow.


The problem

Indie hackers who ship a product need their first customers. Cold outreach to niche micro-influencers works well early. But scraping Instagram is painful and building your own parser is expensive.

I packaged a reliable data source behind a clean API: $0.01 per lead, pay per call, no subscription.


The x402 angle

x402 is an HTTP payment protocol where the client includes a USDC payment token on each API call. The server returns 402 Payment Required with the price, client pays, gets the data back. No OAuth, no billing dashboard, no account.

For AI agents with crypto wallets (increasingly common with Base/USDC adoption), this means:

  1. Agent discovers the API via our MCP server listing
  2. Agent reads the price from the 402 response
  3. Agent pays in USDC and gets leads back β€” zero humans involved

We have an MCP server at api.socialintel.dev/mcp β€” listed on Smithery and Glama β€” so Claude, Cursor, and other MCP-compatible tools can use it natively.


What you can search by:

  • Niche/keyword
  • Location (country, city)
  • Follower count (min)
  • Gender, age range
  • Business accounts only (higher email rate)

Current status:

  • βœ… Live API + MCP server
  • βœ… x402 payments live (USDC on Base)
  • βœ… Listed on Smithery, Glama, awesome-mcp-servers
  • ❌ 0 paid customers β€” just launched, sharing early for feedback

Would love to know:

  • Are you doing cold outreach to influencers? What niches?
  • Is x402 something you want on your own API?
on March 11, 2026
  1. 1

    The no-signup pay-per-call model is the part most people underestimate. Getting AI agents to autonomously pay for API calls sounds clean in theory, but the real friction shows up when the agent hits an unexpected cost and there's no human in the loop to approve it. Worth thinking hard about how you handle spending limits and error states before agents start racking up charges in production.