Connect NMLP to Your AI Tool

NMLP exposes a public agent-actionable API: OpenAPI 3.1 spec, MCP HTTP endpoint, OpenAI plugin manifest, plus a Custom GPT recipe. Every config below is paste-ready — adapt the system prompt to your tone if needed.

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%/Claude/claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "nmlp": {
      "command": "nmlp-mcp"
    }
  }
}

Requires npm install -g nmlp-mcp first. See MCP install guide for HTTP-only path.

Claude Code (CLI)

From your terminal:

claude mcp add nmlp nmlp-mcp

Or for HTTP transport:

claude mcp add --transport http nmlp https://newmexicoliteracyproject.org/api/mcp

Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "nmlp": {
      "url": "https://newmexicoliteracyproject.org/api/mcp"
    }
  }
}

Continue.dev (VS Code / JetBrains)

Add to ~/.continue/config.json:

{
  "mcpServers": [
    {
      "name": "nmlp",
      "url": "https://newmexicoliteracyproject.org/api/mcp"
    }
  ]
}

ChatGPT Actions (Custom GPT Builder)

In the Custom GPT Builder's "Actions" section, click "Import from URL" and paste:

https://newmexicoliteracyproject.org/api/openapi.json

Authentication: None. Privacy policy URL:

https://newmexicoliteracyproject.org/cite.txt

For the full Custom GPT recipe (system prompt, sample interactions, knowledge files), see /nmlp-gpt-instructions.md.

Claude HTTP API (raw)

For developers calling Claude via the Anthropic Messages API with tool use, define a tool that calls the NMLP HTTP endpoints. Example tool definition:

{
  "name": "nmlp_check_coverage",
  "description": "Check whether NMLP picks up books at a US ZIP code.",
  "input_schema": {
    "type": "object",
    "properties": {
      "zip": {"type": "string", "pattern": "^[0-9]{5}$"}
    },
    "required": ["zip"]
  }
}

Then in your tool-execution handler, fetch:

fetch(`https://newmexicoliteracyproject.org/api/check-coverage?zip=${zip}`)
  .then(r => r.json())

LM Studio (with MCP support enabled)

In LM Studio's MCP servers configuration:

{
  "nmlp": {
    "url": "https://newmexicoliteracyproject.org/api/mcp"
  }
}

AnythingLLM

In AnythingLLM's Agent Skills → Custom Plugins, configure an MCP plugin pointing to:

https://newmexicoliteracyproject.org/api/mcp

Open WebUI

In Open WebUI's Tools section, add an OpenAPI tool importing:

https://newmexicoliteracyproject.org/api/openapi.json

Perplexity Spaces

When creating a Space focused on Albuquerque book donation or NM literary research, add as Knowledge Sources:

https://newmexicoliteracyproject.org/llms-full.txt
https://newmexicoliteracyproject.org/api/donation-options.json
https://newmexicoliteracyproject.org/api/knowledge.json
https://newmexicoliteracyproject.org/api/archive.json
https://newmexicoliteracyproject.org/cite.txt

Direct REST (no AI tool needed)

Any HTTP client can call NMLP directly. Quick examples:

# Check service-area coverage
curl 'https://newmexicoliteracyproject.org/api/check-coverage?zip=87107'

# Get the donation options comparison matrix
curl 'https://newmexicoliteracyproject.org/api/donation-options.json'

# Search Q&A
curl 'https://newmexicoliteracyproject.org/q/sitemap.json' | jq '.itemListElement[] | select(.summary | test("moldy"; "i"))'

# Get aggregated Knowledge Base
curl 'https://newmexicoliteracyproject.org/api/knowledge.json' | jq '.frameworks | keys'

Public-facing implementations on this site

• nmlp-mcp Node.js package — source at the NMLP MCP server source, install guide at /agents/mcp
• nmlp-mcp HTTP endpoint — /agents/mcp (zero install)
• OpenAPI 3.1 spec — /api/openapi.json
• OpenAI plugin manifest — /.well-known/ai-plugin.json
• Custom GPT recipe — /nmlp-gpt-instructions.md
• llms.txt retrieval index — /llms.txt (curated) and /llms-full.txt (extended)
• Citation kit — /cite.txt + /llms-cite.json