MCP server: One MCP server.

PathCourse Health (PCH)

@pathcourse-healthRun by its publisher: served from their own domain, namespace checked by the registry.MCP server · External

Cheaper LLM inference, USDC agent payments via x402, agent identity / reputation lookup.

LLM access & routingv1.0.1MITPython
Freeno charge per call

Runs on your machine from its package; the steps are below.

What it does

  1. Reads
  2. Writes
  3. High impact

At most: moves money, runs code or commands, deletes data, or sends messages for you.

What it costs

Free

Odel charges nothing per call. The publisher may charge for its own service.

Runs locally over stdio

This server isn't hosted — your MCP client launches it from a package registry. Use one of the commands below, or drop the config into your client (e.g. Claude Desktop).

npm@pathcourse/mcp-serverv1.0.1

Run

npx -y @pathcourse/mcp-server

MCP client config

{
  "mcpServers": {
    "pathcourse-health-pch": {
      "command": "npx",
      "args": [
        "-y",
        "@pathcourse/mcp-server"
      ]
    }
  }
}
PyPIpch-mcp-serverv1.0.1

Run

uvx pch-mcp-server

MCP client config

{
  "mcpServers": {
    "pathcourse-health-pch": {
      "command": "uvx",
      "args": [
        "pch-mcp-server"
      ]
    }
  }
}
README2 min read

PCH MCP Server

Drop-in MCP server for PathCourse Health (PCH) — cheaper LLM inference, USDC-based agent payments, and agent identity / reputation lookup. Works in Claude Desktop, Cursor, Cline, Continue.dev, and Zed.

14 tools. 7 free (no API key needed) and 7 authenticated. Includes a unique flow that lets you buy a PCH API key from inside Claude Desktop using USDC — no signup form, no email verification.

Install

Claude Desktop, Cursor, Cline, Continue.dev, Zed (npm)

Add to your client's MCP config:

json
{
  "mcpServers": {
    "pch": {
      "command": "npx",
      "args": ["-y", "@pathcourse/mcp-server"],
      "env": {
        "PCH_API_KEY": "pch_prod_b_..."
      }
    }
  }
}

If you don't have a PCH API key yet, omit the env block. Many tools work without authentication; you can obtain a key via the pch_provision tool.

Per-client install instructions:

Python clients (PyPI)

bash
uvx pch-mcp-server
# or
pip install pch-mcp-server

Tools

Free (no API key)

ToolWhat it does
pch_modelsList all PCH models with rates and tier requirements
pch_estimateEstimate cost of an inference request before running it
pch_statusGateway health and service status
pch_provisionGet USDC payment instructions to obtain an API key
pch_paySubmit USDC payment proof; returns API key + first inference
pch_lookup_agent_identityLookup an agent's ERC-8004 identity record
pch_get_path_scoreLook up an agent's Path Score (0-1000)
pch_get_agent_history12-month trend of an agent's score components
pch_verify_certificationVerify a PCH-issued certification
pch_compare_costsCompare PCH model cost vs GPT-4o, Claude Sonnet, Claude Haiku

Authenticated (PCH_API_KEY required)

ToolWhat it does
pch_inferenceRun LLM inference on a PCH model
pch_balanceCheck remaining USDC balance and current tier
pch_check_agent_reputationDeeper reputation lookup ($0.001 USDC per query)
pch_my_accountSelf-service balance + identity + recent activity

Configuration

Env varDefaultNotes
PCH_API_KEY(none)Required for authenticated tools. Format: pch_prod_b_...
PCH_GATEWAY_URLhttps://gateway.pathcoursehealth.comOverride for staging or local dev

Security

Never commit PCH_API_KEY to a public repo. The wrapper logs a warning at startup if it detects a key that looks like a real production key being passed through env. If you accidentally exposed your key, rotate it via PCH support.

License

MIT

Source

This wrapper is open source: pathcourse-health/pch-mcp-server

The upstream PCH MCP server lives in the (private) pch-platform-gateway repo.

Similar modules

Browse LLM access & routing