What is openclaw.net?

This project delivers a self-hosted AI agent runtime built on .NET 10 with NativeAOT support. It runs a gateway that exposes HTTP, WebSocket, and MCP endpoints, plus a browser-based chat and admin console. The runtime includes 48 built-in tools for file operations, web browsing, code execution, and shell access, along with nine channel adapters for messaging platforms like Telegram, Slack, Discord, and email. It works with major LLM providers including OpenAI, Anthropic, Google, Azure, and local models via Ollama. A cross-platform Avalonia desktop companion provides a graphical setup and chat experience with light and dark themes. Deployment options include pre-built desktop bundles for Windows, macOS, and Linux, Docker Compose, or running directly from source with the .NET 10 SDK. The codebase is MIT licensed and developed independently from the upstream OpenClaw project.

Should you self-host openclaw.net?

openclaw.net scores 50/100 on selfhost.directory (how scoring works) and ranks #97 of 209 in AI Agents. It has 528 GitHub stars.

It shipped 8 stable releases in the past year. The latest, v0.4.2, came out on 21 October 2026. New versions land roughly every 3 weeks.

Setup is moderate: expect some configuration beyond a single command. It is MIT-licensed and written mostly in C#.

Its closest alternative on selfhost.directory is Orca (91/100, 108 739 stars). It scores higher overall, so compare the two before you pick. See openclaw.net vs Orca side by side.

Features

48 native tools
9 channel adapters
Admin UI
Browser chat UI
CLI and TUI
Desktop Companion (AgentQi Companion)
Fractal Memory
MCP (Model Context Protocol) endpoint
MetaSkills (DAG workflows)
Multiple LLM provider support
NativeAOT support
OpenAI-compatible HTTP API
Self-hosted AI agent runtime
Tool approvals and observability
Websocket Support

How to install openclaw.net

Install openclaw.net with Docker Compose

Docker and Docker Compose

Environment variables

MODEL_PROVIDER_KEY, OPENCLAW_AUTH_TOKEN, OPENCLAW_MODEL, MODEL_PROVIDER_ENDPOINT, OPENCLAW_DOMAIN, OPENCLAW_WORKSPACE

version: "3.8"

services: openclaw: image: ${OPENCLAW_IMAGE:-openclaw.net:local} build: context: . dockerfile: Dockerfile container_name: openclaw-gateway restart: unless-stopped ports: - "18789:18789" environment: # Required: set your LLM API key - MODEL_PROVIDER_KEY=${MODEL_PROVIDER_KEY:?Set MODEL_PROVIDER_KEY} # Required for non-loopback bind - OPENCLAW_AUTH_TOKEN=${OPENCLAW_AUTH_TOKEN:?Set OPENCLAW_AUTH_TOKEN} # Optional overrides - MODEL_PROVIDER_MODEL=${OPENCLAW_MODEL:-gpt-4o} - MODEL_PROVIDER_ENDPOINT=${MODEL_PROVIDER_ENDPOINT:-} # ASP.NET Core config via env (double underscore = section separator) - OpenClaw__BindAddress=0.0.0.0 - OpenClaw__Port=18789 - OpenClaw__Tooling__AllowShell=false - OpenClaw__Tooling__AllowedReadRoots__0=/app/workspace - OpenClaw__Tooling__AllowedWriteRoots__0=/app/workspace # Safe default for public binds: disable JS plugin bridge unless explicitly enabled. - OpenClaw__Plugins__Enabled=false # If running behind a reverse proxy (Caddy/nginx), you can enable forwarded headers: # - OpenClaw__Security__TrustForwardedHeaders=true # - OpenClaw__Security__KnownProxies__0= volumes: # Persist memory/session data - openclaw-memory:/app/memory # Optional: mount a workspace for file tools - ${OPENCLAW_WORKSPACE:-./workspace}:/app/workspace healthcheck: test: ["CMD", "/app/OpenClaw.Gateway", "--health-check"] interval: 30s timeout: 5s retries: 3 start_period: 10s

Optional: reverse proxy with automatic TLS

caddy: image: caddy:2-alpine container_name: openclaw-caddy restart: unless-stopped ports: - "443:443" - "80:80" environment: - DOMAIN=${OPENCLAW_DOMAIN:-localhost} volumes: - ./deploy/Caddyfile:/etc/caddy/Caddyfile:ro - caddy-data:/data - caddy-config:/config depends_on: openclaw: condition: service_healthy profiles: - with-tls

volumes: openclaw-memory: caddy-data: caddy-config:

Requires MODEL_PROVIDER_KEY and OPENCLAW_AUTH_TOKEN environment variables. Optional Caddy reverse proxy with TLS via 'with-tls' profile.

Install openclaw.net with Windows

Extract the archive and launch Companion from the companion folder.
Open Setup. In the redesigned Companion, you can also choose Set up my assistant in Chat.
Choose your provider and model. Enter a hosted provider key, or choose Ollama for a local model server.
Click Set Up and Start.

Available for Windows x64, macOS Apple Silicon, and Linux x64. Windows and macOS release archives are currently unsigned.

Install openclaw.net with Build from Source

Environment variables

MODEL_PROVIDER_KEY

git clone https://github.com/clawdotnet/openclaw.net cd openclaw.net export MODEL_PROVIDER_KEY="sk-..." dotnet run --project src/OpenClaw.Cli -c Release -- start

Requires Git and .NET 10 SDK. Optional: Node.js 20+ for upstream-style TS/JS plugin support.