Caution
This repository is public. Never commit secrets, API keys, tokens, .env files, or any personal configuration here.
A personal sandbox for learning OpenClaw — a free, open-source autonomous AI agent (formerly Clawdbot / Moltbot) that runs locally and is reachable via the messaging platforms you already use (Telegram, Discord, Slack, WhatsApp, Signal, and many more).
OpenClaw connects to a large language model of your choice (Claude, GPT-4o, Gemini, DeepSeek, …) and acts as a personal assistant you interact with through chat. It can browse the web, manage files, summarise documents, handle emails, schedule calendar entries, and more — all running on your own hardware.
Upstream project: https://github.com/openclaw/openclaw
Experiments with running OpenClaw in a Docker Compose sandbox. Nothing here is meant for production.
compose.yaml # Docker Compose stack (openclaw-gateway + openclaw-cli)
.env.example # Template — copy to .env and fill in your secrets
Makefile # Convenience targets
.gitignore # Keeps .env and runtime state out of git
# 1. Copy the env template and fill in at least one LLM key and one channel token
cp .env.example .env
# 2. Pull the image and start the gateway
make pull
make up
# 3. Open the web UI
open http://localhost:18789
# 4. Or drop into an interactive CLI session
make cli| Target | What it does |
|---|---|
make up |
Start the gateway in detached mode |
make down |
Stop and remove containers |
make logs |
Follow gateway logs |
make pull |
Pull the latest image |
make cli |
Open an interactive CLI session |
Runtime state (config, workspace, auth secrets) is stored in .openclaw/ and .openclaw-auth/ — both git-ignored.