obot

Obot

Obot is an open-source MCP Gateway and AI platform that can be deployed in the cloud or on-prem.

Getting Started

To quickly try a live demo of the Obot MCP Gateway and chat experience, visit https://chat.obot.ai.

To run Obot yourself, you’ll need to setup Docker with something like Docker Desktop. Once that’s ready, run:

docker run -d --name obot -p 8080:8080 -v /var/run/docker.sock:/var/run/docker.sock -e OPENAI_API_KEY=<API KEY> ghcr.io/obot-platform/obot:latest

Then open your browser to http://localhost:8080 to access the Obot UI.

[!TIP] You need to replace <API KEY> with your OpenAI API Key.

Setting this is optional, but you'll need to setup a model provider from the Admin UI before using chat.

You can also set ANTHROPIC_API_KEY here as well, setting the value to your Anthropic API Key.

Setting both is also supported, but OpenAI models will be set as the defaults.

For more installation methods, see our Installation Guide.

The Three Parts of Obot

The platform consists of three main components that work together to deliver a comprehensive AI solution.

🔌 MCP Gateway

The MCP Gateway is where users discover and connect to MCP servers using any MCP client. It provides:

• Server Discovery – Browse a catalog of MCP servers tailored to your role and permissions
• Configuration Management – Manage all MCP server settings and credentials in one place
• Upgrade Management – Receive notifications about available server upgrades and apply them easily
• Broad Client Support – Connect with local clients such as Claude Desktop and VS Code or use our hosted Obot Chat
• OAuth 2.1 Authentication – Securely authenticate with external services

🗣️ Chat

The Chat Interface is where users interact with AI through natural, conversational chat. It’s the primary way to ask questions, get answers, and work with connected tools and data. Key features include:

• Chat Threads – Keep discussions organized and maintain context over time
• MCP Server Integration – Connect to SaaS platforms, APIs, and other tools through MCP servers
• Knowledge Integration – Use built-in RAG to add relevant knowledge to your conversations
• Tasks - Create and schedule repeatable tasks that can leverage all the same capabilities as Chat
• Project-Based Customization – Tailor AI's behavior to meet your needs with custom instructions, knowledge, and MCP servers at the project level

⚙️ Admin

The Admin Interface provides comprehensive platform management tools for administrators:

• Catalog Management – Create and update MCP server entries using GitOps or the admin portal
• Server Deployment and Hosting - Let Obot deploy and host MCP servers to ease your operational burden
• Access Control Rules – Define which users and groups can access specific MCP servers
• Audit Logging – Track and record all MCP server and client interactions
• Request Filtering – Programmatically inspect and reject requests to/from MCP servers for enhanced security and compliance
• User Management – Manage users, groups, and access permissions
• Model Provider Management – Configure and manage LLM providers and settings for the Chat Interface
• Centralized Authentication - Integrate with your existing auth provider to ensure proper user authentication and authorization
• Monitoring – View system health metrics and usage analytics

How They Work Together

These three components create a powerful, integrated AI platform:

1. Users interact with Obot projects through the Chat Interface and MCP Servers through the MCP Gateway.
2. Users and MCP Clients leverage tools via the MCP Gateway
3. Administrators manage the entire platform through the Admin Interface

Key Features

• Self-Hosted: Deploy on your own infrastructure for complete control
• MCP Standard: Built on the open Model Context Protocol for maximum interoperability
• Enterprise Security: OAuth 2.1 authentication, encryption, and audit logging
• Extensible: Easy integration with custom tools and services

Next Steps

For more information checkout our Docs.