Model Context Protocol (MCP) Server for Graphlit Platform

Name: Model Context Protocol (MCP) Server for Graphlit Platform
Author: graphlit

@graphlit

リポジトリホームページ

Model Context Protocol (MCP) Server for Graphlit Platform について

#graphlit #mcp-server #integration

基本情報

カテゴリ

その他

ライセンス

MIT

ランタイム

node

トランスポート

stdio

公開者

graphlit

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

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

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Model Context Protocol (MCP) Server for Graphlit Platform?

The Graphlit MCP Server enables integration between MCP clients (such as Cursor, Windsurf, Goose, or Cline) and the Graphlit platform. It ingests content from Slack, Discord, websites, Google Drive, email, Jira, Linear, GitHub, and more into a Graphlit project, which then acts as a searchable, RAG-ready knowledge base. It is built for developers who need to centralize knowledge from multiple tools and retrieve it from within an MCP client.

How to use Model Context Protocol (MCP) Server for Graphlit Platform?

Install the server using npx -y graphlit-mcp-server and configure it with three required environment variables: GRAPHLIT_ENVIRONMENT_ID, GRAPHLIT_ORGANIZATION_ID, and GRAPHLIT_JWT_SECRET (found in the Graphlit API settings dashboard). Optional credentials for data connectors (Slack, Google, Notion, etc.) can be added. One-click install is available for VS Code; manual configuration is supported via JSON blocks for VS Code, Windsurf, Cline, Cursor, or any MCP client.

Key features of Model Context Protocol (MCP) Server for Graphlit Platform

Ingest files, web pages, messages, posts, emails, issues, and text.
Built-in web crawling and web search (no separate tools needed).
RAG via “Prompt LLM Conversation” tool.
Extract structured JSON from text.
Publish audio (ElevenLabs) and images (OpenAI).
Documents are extracted to Markdown; audio/video are transcribed.

Use cases of Model Context Protocol (MCP) Server for Graphlit Platform

Create a unified, searchable knowledge base from Slack, email, Jira, GitHub, and other tools.
Automate ingestion of web pages, PDFs, and other documents for later retrieval.
Enable RAG workflows by querying ingested content directly from an MCP client.
Crawl websites and search the web without additional integrations like Firecrawl or Exa.
Manage collections, feeds, and conversations for organizing knowledge.

FAQ from Model Context Protocol (MCP) Server for Graphlit Platform

What are the prerequisites for using the Graphlit MCP Server?

You need Node.js (version 18.x or higher) and an active Graphlit Platform account with access to the API settings dashboard.

What data connectors does the server support?

It supports Microsoft Outlook, Google Mail, Notion, Reddit, Linear, Jira, GitHub Issues, Google Drive, OneDrive, SharePoint, Dropbox, Box, GitHub, Slack, Microsoft Teams, Discord, Twitter/X, and Podcasts (RSS).

How do I install the server in VS Code?

You can use the one-click install buttons provided, or manually add a JSON block to your User Settings (JSON) or .vscode/mcp.json file. The JSON must include the npx command and the three required environment variables.

Can I use additional credentials for data connectors?

Yes. Only the three Graphlit environment variables are required, but you can optionally set tokens for Slack, Discord, Twitter, Google Email, Linear, GitHub, Jira, and Notion in the env section of the configuration.