Repolode
Submit
clawdotnet

clawdotnet/openclaw.net

A self-hosted AI agent runtime in .NET with a desktop companion app, supporting multiple model providers, tools, memory, and chat channels.

View on GitHubWebsite
Stars
527
Forks
105
Open issues
2
Last commit
yesterday

What is it?

OpenClaw.NET is a self-hosted runtime for running AI agents on your own machine or a team server. It connects to hosted models such as OpenAI, Claude, Gemini and Azure OpenAI, or to local models through Ollama and optional embedded inference, and exposes the agent through a browser chat, a CLI, OpenAI-compatible HTTP endpoints, MCP and WebSockets.

The runtime includes streaming, tool calling with approvals, sessions, memory and token-usage reporting. A large set of native tools covers files, web access, databases and email, and channel adapters cover Telegram, WhatsApp, Slack, Discord, Teams and others. Extensions come through MCP servers, SKILL.md packages and reusable DAG workflows called MetaSkills.

AgentQi Companion is an Avalonia desktop app that handles setup, connections, memory and settings in one window. The gateway binds to 127.0.0.1 by default and includes URL safety checks for outbound fetches. Developers can build on it in .NET and deploy supported parts as NativeAOT binaries.

Key features

  • 01Connects to OpenAI, Claude, Gemini, Azure OpenAI, DeepSeek, Ollama and OpenAI-compatible providers.
  • 02Provides a browser chat, admin UI, CLI, terminal UI and desktop companion app.
  • 03Offers OpenAI-compatible HTTP endpoints, MCP and WebSocket access.
  • 04Includes 80+ native tool surfaces and channel adapters for common messaging platforms.
  • 05Supports sessions, memory, tool approvals and token-usage reporting.
  • 06Can be deployed as NativeAOT binaries for supported features.

Use cases

  • Running a private AI assistant on your own machine with a local or hosted model.
  • Hosting a shared AI gateway for a team with authentication and tool restrictions.
  • Connecting an agent to messaging channels such as Telegram or Slack.
  • Building custom .NET agent tools and integrating them through MCP.

Topics

#ai-agent#dotnet#self-hosted#llm#tool-calling#mcp#csharp#personal-ai#agent-harness#agent-runtime

FAQ

What is openclaw.net?
A self-hosted AI agent runtime in .NET with a desktop companion app, supporting multiple model providers, tools, memory, and chat channels. OpenClaw.NET is a self-hosted runtime for running AI agents on your own machine or a team server. It connects to hosted models such as OpenAI, Claude, Gemini and Azure OpenAI, or to local models through Ollama and optional embedded inference, and exposes the agent through a browser chat, a CLI, OpenAI-compatible HTTP endpoints, MCP and WebSockets.
Is openclaw.net free to use? What license does it have?
openclaw.net is open source under the MIT license. Check the license terms before using it in commercial products.
Is openclaw.net still maintained?
Actively maintained. The last commit to openclaw.net was on Oct 21, 2026.
What can I use openclaw.net for?
Running a private AI assistant on your own machine with a local or hosted model. Hosting a shared AI gateway for a team with authentication and tool restrictions. Connecting an agent to messaging channels such as Telegram or Slack. Building custom .NET agent tools and integrating them through MCP.
Where can I get openclaw.net?
The source code and installation instructions are on GitHub: https://github.com/clawdotnet/openclaw.net

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects