Skip to main content
Glama

speclib-mcp

MCP server for searching, retrieving, and creating SpecLib specs. Allows AI agents to discover, read, and write specifications via the Model Context Protocol.

Tools

Tool

Description

search_specs

Search specs by query, scope, and/or content type

get_spec

Get full spec content by ID or scope/slug

list_scopes

List all available scopes

get_recipe

Get a recipe with its bundled specs

create_spec

Create a new spec (requires API token)

update_spec

Update an existing spec (requires API token)

Resources

URI

Description

spec://{scope}/{slug}

Read a spec as markdown

Usage

One-liner

npx github:rang501/speclib-mcp

Claude Code

Add to your MCP settings (.mcp.json or project config):

{
  "mcpServers": {
    "speclib": {
      "command": "npx",
      "args": ["-y", "github:rang501/speclib-mcp"],
      "env": {
        "SPECLIB_API_URL": "http://localhost:3000",
        "SPECLIB_API_TOKEN": "your-api-token"
      }
    }
  }
}

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "speclib": {
      "command": "npx",
      "args": ["-y", "github:rang501/speclib-mcp"],
      "env": {
        "SPECLIB_API_URL": "http://localhost:3000",
        "SPECLIB_API_TOKEN": "your-api-token"
      }
    }
  }
}

Local development

git clone https://github.com/rang501/speclib-mcp.git
cd speclib-mcp
npm install
node index.mjs

Configuration

Environment Variable

Default

Description

SPECLIB_API_URL

http://localhost:3000

SpecLib instance URL

SPECLIB_API_TOKEN

(none)

API token for write operations (generate in Account > API Tokens)

Authentication

Reading public specs works without authentication. To create or update specs, you need an API token:

  1. Log in to your SpecLib instance

  2. Go to Account > API Tokens

  3. Create a new token and copy it

  4. Set it as SPECLIB_API_TOKEN in your MCP client config

Tokens that have not been used for over a year are automatically expired.

Requirements

  • Node.js >= 20

  • A running SpecLib instance