mcp-chat

<a href="https://chat.pipedream.com/"> <img alt="MCP Chat by Pipedream" src="app/(chat)/opengraph-image.png"> <h1 align="center">MCP Chat by Pipedream</h1> </a> <p align="center"> MCP Chat is a free, open-source chat app built using the AI SDK, and Pipedream MCP, which provides access to nearly 3,000 APIs and more than 10,000 tools. Use this as a reference to build powerful AI chat applications. </p> <p align="center"> <a href="#features"><strong>Features</strong></a> · <a href="#model-providers"><strong>Model Providers</strong></a> · <a href="#prerequisites"><strong>Prerequisites</strong></a> · <a href="#deploy-your-own"><strong>Deploy Your Own</strong></a> · <a href="#running-locally"><strong>Running Locally</strong></a> </p> <br/> > **Check out the app in production at [chat.pipedream.com](https://chat.pipedream.com) and refer to [Pipedream's developer docs](https://pipedream.com/docs/connect/mcp/developers) for the most up to date information.** ## Features - **MCP integrations**: Connect to thousands of APIs through Pipedream's MCP server with built-in auth - **Automatic tool discovery**: Execute tool calls across different APIs via chat - **Uses the [AI SDK](https://sdk.vercel.ai/docs)**: Unified API for generating text, structured objects, and tool calls with LLMs - **Flexible LLM and framework support**: Works with any LLM provider or framework - **Data persistence**: Uses [Neon Serverless Postgres](https://vercel.com/marketplace/neon) for saving chat history and user data and [Auth.js](https://authjs.dev) for simple and secure sign-in ## Model Providers The demo app currently uses models from Anthropic, OpenAI, and Gemini, but the AI SDK supports [many more](https://sdk.vercel.ai/providers/ai-sdk-providers). ### Prerequisites To run or deploy this app, you'll need: 1. A [Pipedream account](https://pipedream.com/auth/signup) 2. A [Pipedream project](https://pipedream.com/docs/projects/#creating-projects). Accounts connected via MCP will be stored here. 3. [Pipedream OAuth credentials](https://pipedream.com/docs/rest-api/auth/#oauth) 4. An [OpenAI API key](https://platform.openai.com/api-keys) ## Deploy Your Own One-click deploy this app to Vercel: [](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FPipedreamHQ%2Fmcp-chat&env=PIPEDREAM_CLIENT_ID,PIPEDREAM_CLIENT_SECRET,PIPEDREAM_PROJECT_ID,PIPEDREAM_PROJECT_ENVIRONMENT,AUTH_SECRET,GOOGLE_CLIENT_ID,GOOGLE_CLIENT_SECRET,OPENAI_API_KEY,EXA_API_KEY,POSTGRES_URL&envDescription=API%20keys%20need%20to%20run%20the%20app) ## Running locally 1. Copy the environment file and add your credentials: ```bash cp .env.example .env # Edit with your values ``` Note that for easier development, chat persistence and application sign-in are disabled by default in the `.env.example` file: ```bash # In your .env file DISABLE_AUTH=true DISABLE_PERSISTENCE=true ``` 2. Install dependencies and start the app: We recommend using [asdf](https://asdf-vm.com/) to manage core deps like Node. Install it and run ```bash asdf install ``` Then: ```bash pnpm install pnpm dev ``` Your local app should now be running on [http://localhost:3000](http://localhost:3000/) 🎉 ### Enabling chat persistence 1. Run all required local services: ```bash docker compose up -d ``` 2. Run migrations: ```bash POSTGRES_URL=postgresql://postgres@localhost:5432/postgres pnpm db:migrate ```