Obot, developed by Dawosch, is an open-source platform for managing, securing, and governing AI ecosystems within enterprise environments. Distributed as Docker containers, Kubernetes workloads, and a Go binary, it acts as an AI Control Plane and MCP Gateway. Obot connects AI clients to models and tools, providing shared infrastructure. Its backend is written in Go 1.26.0, ensuring a performant foundation for its operations. This mcp category solution helps organizations control their AI integrations.

What does Obot connect to?

Obot integrates with various external services to manage AI operations. It connects to Model Context Protocol (MCP) servers, Large Language Model (LLM) providers, and S3-compatible object storage. For version control, it links with Git providers such as GitHub and GitLab. Furthermore, Obot supports multiple authentication providers, including GitHub, Google, and Okta, to secure access. Its role as an an AI Control Plane and MCP Gateway establishes it as a central hub for AI client connections to models and tools.

Core features and tools

Obot provides several callable tools and components. The MCP Gateway acts as a single, governed entry point for MCP servers, supporting proxying, creating composite MCP servers, and access control. The Obot CLI enables users and AI clients to discover, install, and manage approved MCP servers and skills, with the "obot mcp search" command for finding servers. Obot Sentry extends governance to user devices by inventorying AI clients, servers, skills, and plugins, and recording local tool calls. An LLM Gateway serves as a model endpoint compatible with OpenAI and Anthropic APIs.

Who is it for?

Obot is designed for platform engineers, IT architects, and DevOps teams aiming to implement a governed, enterprise-grade AI platform. It also serves developers and business users who need to discover and utilize AI tools within a controlled environment. Confirmed compatible AI clients include desktop agents and tools such as Claude Code, Codex, Cursor, VS Code, and other IDEs and CLIs. This is not the right fit if you require a simple, single-user AI tool without enterprise governance needs.

How to get started

Access is granted through a standard OAuth authorization flow, requiring clients to first authenticate with Obot via a configured identity provider. Setup involves connecting Obot to an identity provider of choice, such as GitHub, Google, or Okta.

Use cases

• Securely connecting AI agents to company data: Obot's MCP Gateway secures connections with SSO and OIDC, enabling AI agents to interact with enterprise systems.
• Managing and governing MCP servers and skills: Organizations deploy and configure MCP server types, controlling user access via registries and policies, with activity monitoring through audit logs.
• Auditing and enforcing policies on AI activity: Obot captures tool calls and LLM requests, enforcing policies on agent actions and extending governance to local AI activity via Obot Sentry.

Obot offers an effective platform for enterprise AI governance, providing a centralized control plane for managing AI ecosystems. It secures connections, manages AI assets, and enforces policies across various integrations and user devices. Its modular design and support for standard protocols make it suitable for complex enterprise environments. Dawosch's Obot addresses the need for controlled AI deployment and usage.