agent402-mcp

MCP server for Agent402 — 1,100+ pay-per-call web tools for AI agents, paid per call in USDC via the x402 protocol, or with compute (proof-of-work) when no wallet is configured. Built by Mikey Petrillo.

Your agent gets browser rendering, screenshots, PDF text extraction, URL→markdown, live web search + web answers with citations, live financial/crypto/macro data (Yahoo stock quotes, CoinGecko, FRED, ECB FX, World Bank, yield curve), SEC EDGAR filings (10-K/10-Q text, XBRL, insider, 13F, IPO calendar), DNS/TLS/WHOIS, wallet-keyed shared memory, and ~1,040 utility/conversion tools — with payment handled invisibly underneath the MCP calls. No signup, no API key.

Quick start

Zero install (hosted connector): add https://agent402.tools/mcp as a remote MCP server — e.g. claude.ai → Settings → Connectors → Add custom connector. The pure-CPU tools run free there (rate-limited); for the full catalog and no rate limit, run this package locally with a wallet:

With a funded wallet (USDC on Base) — every tool available:

{
  "mcpServers": {
    "agent402": {
      "command": "npx",
      "args": ["-y", "agent402-mcp"],
      "env": { "AGENT_KEY": "0xYOUR_PRIVATE_KEY" }
    }
  }
}

Without a wallet — the ~1000 pure-CPU tools work free via proof-of-work (the network/browser/memory tools will ask for a wallet):

{
  "mcpServers": {
    "agent402": { "command": "npx", "args": ["-y", "agent402-mcp"] }
  }
}

Claude Code: claude mcp add agent402 -- npx -y agent402-mcp

How it works

• On startup the server reads the live catalog from https://agent402.tools/api/pricing + /openapi.json.
• The high-value tools (extract, render, screenshot, pdf, meta, dns, http-check, tls-cert, whois, the memory-* coordination tools, hash) are exposed as first-class MCP tools.
• The other ~1000 tools are reachable via search_tools (find by description) + call_tool (call by slug) — keeping your context window small.
• When a call hits HTTP 402: with AGENT_KEY set, the server signs an x402 USDC payment and retries; without a key it solves the tool's proof-of-work challenge (~0.2 s of CPU) on the eligible tools.
• payment_info tells the model which mode it's in and what a wallet would unlock.

Workflows (skill packs)

For jobs that no single tool covers (e.g. "audit a domain", "build a stock brief"), Agent402 ships curated multi-tool skill packs. They're surfaced as standard MCP prompts, so any MCP-aware client picks them up automatically:

• prompts/list returns each pack with typed arguments.
• prompts/get { name: "<slug>", arguments: { … } } returns the rendered task template — a Claude-ready plan with the chosen tools wired in.
• search_tools also surfaces matching workflows alongside individual tools, so a task-shaped query points the agent at the right plan, not just the raw tools.

Configuration

| env | default | meaning | | --- | --- | --- | | AGENT_KEY | (unset) | Hex private key of a wallet funded with USDC on Base. Unset = proof-of-work mode. | | AGENT402_URL | https://agent402.tools | Target service (point at your own deployment). | | AGENT402_TOOLS | curated set | Comma-separated slugs to expose as first-class tools. | | AGENT402_MAX_PER_CALL | unlimited | Refuse any single call priced above this many USD (e.g. 0.01). | | AGENT402_BUDGET | unlimited | Hard cap on total USDC spent per session (e.g. 1.00). |

Spend controls are enforced before a payment is signed — a runaway model is refused, not billed. payment_info reports the caps, what's been spent, and what remains. Use a dedicated low-value wallet for AGENT_KEY, funded only with what you intend to spend — calls cost $0.001–$0.02 each.

Test

From the repo root: node mcp/test.js (boots a local paywalled instance and drives the MCP server with a real client; the proof-of-work path settles real challenges).