LLM Kit

The FastComments LLM Kit helps AI coding assistants like Claude, ChatGPT, and Cursor understand and integrate with FastComments by providing them access to our complete documentation through a simple search API.

By adding a snippet to your project's AGENT.md or CLAUDE.md file, you enable AI assistants to quickly search and retrieve relevant FastComments documentation, making implementation faster and more accurate.

Get the LLM Kit

Add this snippet to your project's AI assistant configuration files (like AGENT.md or CLAUDE.md) that teaches AI coding assistants how to search and access FastComments documentation.

FastComments LLM Kit

Copy

## FastComments Integration
FastComments is an embeddable live commenting platform with libraries for many frontend and backend integrations, along with client and server-side SDKs and APIs.
When working with FastComments, you can search the FastComments documentation by calling this API:
https://docs-search.fastcomments.com/search?query=<search_query>&full=true&tenantId=demo
Replace `<search_query>` with your search terms (URL encoded). The API returns relevant documentation snippets to help with implementation.

This way your AI assistant can easily get up-to-date information and provide more accurate implementations in less time.

Usage Example

When using an AI assistant like Claude Code or Cursor, you can ask questions like:

"How do I install FastComments on React?"
"Show me how to configure SSO with FastComments"
"What are the FastComments API authentication options?"

The AI assistant will automatically search the documentation using the provided API endpoint and give you accurate, up-to-date answers based on the official documentation.

MCP Server

FastComments runs a hosted Model Context Protocol (MCP) server so AI coding assistants and agentic clients can call the FastComments API directly. Every tool the MCP server exposes is auto-generated from the public OpenAPI spec, so anything the REST API can do, an MCP client can do.

The endpoint is stateless and streamable-HTTP based. There's no session to keep alive, no client registration step, and no server-side state per client.

Endpoint

MCP Endpoint

Copy

https://fastcomments.com/mcp?tenantId=YOUR_TENANT_ID&API_KEY=YOUR_API_KEY

Authentication uses the same API key as the REST API. You can also pass tenantId and the key as x-tenant-id and x-api-key HTTP headers if your client supports custom headers.

Pre-filled setup

The dashboard has a setup helper that generates the URL and ready-to-paste config snippets for popular MCP clients. Go to your account dashboard and open Integrate -> MCP Server, or visit it directly:

Setup Page

Copy

https://fastcomments.com/auth/my-account/mcp-setup

Pick which API key to use from the dropdown, then copy any of the generated snippets.

Claude Code

Claude Code Setup

Copy

claude mcp add --transport http fastcomments 'https://fastcomments.com/mcp?tenantId=YOUR_TENANT_ID&API_KEY=YOUR_API_KEY'

After it's registered, run /mcp inside a Claude Code session to confirm the connection and list available tools.

Claude Desktop / Cursor

Add this block to your client's MCP servers config (claude_desktop_config.json for Claude Desktop, mcp.json for Cursor):

MCP Client Config

Copy

{
  "mcpServers": {
    "fastcomments": {
      "type": "http",
      "url": "https://fastcomments.com/mcp?tenantId=YOUR_TENANT_ID&API_KEY=YOUR_API_KEY"
    }
  }
}

Security

The API key is embedded in the URL. Treat the URL like a secret: do not paste it into public chats, screenshots, or commits. If a key is exposed, rotate it on the API Keys page in your dashboard.

Language 🇺🇸 English (US)

LLM Integration

MCP Server

LLM Kit

Get the LLM Kit

Usage Example

MCP Server

Endpoint

Pre-filled setup

Claude Code

Claude Desktop / Cursor

Security