MCP ExplorerExplorer

Summon App

@TrySummonon 10 days ago
12 Apache-2.0
FreeCommunity
AI Systems
#agent#agents#llm#mcp
Your All-in-One Desktop App for Building, Testing and Sharing Model Context Protocol (MCP) Servers & AI Agents.

Overview

What is Summon App

Summon-app is an all-in-one desktop application designed for building, testing, and sharing Model Context Protocol (MCP) servers and AI agents.

Use cases

Use cases for summon-app include creating a unified MCP server from multiple APIs, testing AI interactions in real-time, and securely managing API credentials for various applications.

How to use

To use summon-app, download and install the application, import OpenAPI specifications, configure your MCP server settings, and utilize the AI Playground for testing. You can also connect to existing MCP servers and manage API authentication.

Key features

Key features include OpenAPI to MCP generation, API authentication configuration, external MCP support, an interactive AI Playground, live tool editing, and secure credential storage.

Where to use

Summon-app can be used in software development, API management, AI model testing, and any scenario that requires building and managing MCP servers.

Content

Summon Logo

Latest Release
License
Y Combinator S25

The Collaborative Desktop App for Teams to Build, Test, and Manage MCP Servers & AI Agents.

Download the latest release


The main features of the Summon app are:

Features Description
OpenAPI to MCP Generation Import OpenAPI specs, visually select endpoints to include in your MCP server. Combine multiple APIs into a single, cohesive server.
API Auth Configure authentication for real API calls or use mock servers to test your MCP tools without external dependencies.
External MCP Support Connect to any existing MCP server (local or remote) via mcp.json configuration. Support for CLI-based servers with automatic tool discovery.
AI Playground Test MCP tool calling in an interactive chat interface with multiple AI models including OpenAI, Anthropic and more.
Live Tool Editing Modify tool names, descriptions, and input schemas on the fly without restarting servers. View diffs of your modifications in real-time.
Datasets Create and manage test datasets with conversation scenarios, expected tool calls, and evaluation criteria. AI-assisted generation of test cases.
Evals Run evaluations of your MCPs against dataset items. Track performance metrics and tool calling accuracy.

📺 Showcase

https://github.com/user-attachments/assets/f3ff8ef4-2ec8-4aa3-9d30-8d10bdc0b6ea

🚀 Getting Started

1. Download

Grab the latest release for your operating system.

2. Install

Follow the standard installation procedure for your OS:

  • macOS: Open the .dmg file and drag Summon.app to your Applications folder.
  • Windows: Run the .exe installer.
  • Linux: Use the .rpm or .deb package.

3. Launch Summon

[!NOTE]
Keychain Access: You may be prompted to grant keychain access when first using Summon. This is because Summon encrypts and stores API keys locally on your device using your system’s secure keychain services. All credentials are stored locally and never sent to external servers.

Open the application and you’re ready to go!

🗺️ Roadmap

We have exciting plans for Summon! Our collaborative roadmap, including upcoming features and long-term vision, can be found on our website.

💻 For Developers (Building from Source)

If you want to contribute or run Summon from source:

  1. Clone the repository:

    git clone https://github.com/TrySummon/summon-app.git
    cd summon-app
    
  2. Install dependencies:

    npm install
    
  3. Run the development server:

    npm start
    
  4. Build the application:

    npm run make # For your current platform
    

    (See package.json for other build scripts.)

🏗️ Technology Stack

  • Electron: For cross-platform desktop application framework.
  • Tailwind & shadcn: For the user interface.
  • TanStack Router & Query: For routing and data fetching/caching.
  • Zustand: For global state management.
  • Model Context Protocol SDK: For MCP client/server interactions.
  • AI SDK (Vercel AI SDK): For LLM interactions in the Playground.
  • Express: For running local MCP servers and API mock servers.

❤️ Contributing

Contributions are welcome! Whether it’s bug reports, feature requests, or code contributions, please feel free to:

  1. Check our Contribution Guidelines.
  2. Open an issue on the GitHub Issues page.
  3. Fork the repository and submit a pull request.

Made with ❤️ by the Summon Team

Tools

No tools

Comments