Guide: OpenClaw with Daytona — Up and Running in 2 Minutes

guides/20260531_openclaw_daytona_2min.md

@@ -0,0 +1,211 @@
++++
+tags = ['openclaw', 'daytona', 'ai-development', 'getting-started', 'tutorial']
+---
+
+# OpenClaw with Daytona: Up and Running in 2 Minutes
+
+## Introduction
+
+OpenClaw and Daytona are a natural pairing. OpenClaw gives you a self-hosted AI agent that can automate your workflows. Daytona gives you instant, reproducible development environments. Together, you can go from zero to a fully configured agent environment in under two minutes — no manual dependency installation, no "works on my machine" drama.
+
+This guide walks through the fastest path to a running OpenClaw + Daytona setup.
+
+## Prerequisites
+
+- A machine running macOS, Linux, or Windows (via WSL2)
+- Docker installed and running
+- A Git provider account (GitHub, GitLab, or Bitbucket)
+- ~2 GB of free disk space
+
+## Step 1: Install Daytona
+
+**macOS/Linux:**
+```bash
+curl -fsSL https://get.daytona.io | bash
+```
+
+**Verify:**
+```bash
+daytona version
+```
+
+**Start the server:**
+```bash
+daytona serve
+```
+
+The Daytona server starts on `localhost:10001` by default.
+
+## Step 2: Create Your OpenClaw Workspace
+
+Create a `.devcontainer/devcontainer.json` in your project:
+
+```json
+{
+  "name": "OpenClaw Workspace",
+  "image": "mcr.microsoft.com/devcontainers/base:ubuntu",
+  "features": {
+    "ghcr.io/devcontainers/features/docker-in-docker:2": {},
+    "ghcr.io/devcontainers/features/node:1": {}
+  },
+  "forwardPorts": [3000, 10001],
+  "postCreateCommand": "npm install -g @anthropic/anthropic-ai",
+  "customizations": {
+    "vscode": {
+      "extensions": ["anthropic.claude-code"]
+    }
+  }
+}
+```
+
+This gives you:
+- Ubuntu base with Docker-in-Docker support
+- Node.js for running OpenClaw
+- Claude Code VS Code extension pre-installed
+- Port forwarding for the OpenClaw gateway (3000) and Daytona server (10001)
+
+## Step 3: Spin Up the Environment
+
+```bash
+daytona create .
+```
+
+Daytona reads the devcontainer config, pulls the image, installs features, and runs the post-create command. Typical spin-up time: **30–90 seconds**.
+
+## Step 4: Connect Your IDE
+
+**VS Code:**
+1. Open VS Code
+2. Install the Daytona extension (search "Daytona" in Extensions)
+3. Click the Daytona icon in the sidebar
+4. Select your environment and click "Connect"
+
+**Terminal:**
+```bash
+daytona ssh <env-name>
+```
+
+## Step 5: Initialize OpenClaw
+
+Inside the Daytona environment:
+
+```bash
+# Clone the OpenClaw repo
+git clone https://github.com/openclaw/openclaw.git
+cd openclaw
+
+# Run the onboard wizard
+openclaw onboard
+```
+
+The onboard wizard walks through:
+- Gateway configuration
+- Provider/model routing (choose your LLM provider)
+- Channel setup (Telegram, Discord, iMessage, etc.)
+- Skill installation
+- Workspace structure
+
+## Step 6: Verify the Gateway
+
+```bash
+openclaw gateway status
+```
+
+You should see:
+```
+Gateway: running
+Port: 3000
+Channels: configured
+Skills: loaded
+```
+
+## What You Have Now
+
+- **OpenClaw gateway** running inside a Daytona-managed environment
+- **Docker-in-Docker** support for running containers from within your agent
+- **Persistent workspace** — your OpenClaw config is version-controlled in the devcontainer
+- **Reproducible** — any team member gets the same setup with `daytona create`
+
+## Common Configurations
+
+### Self-Hosted LLM
+
+Point OpenClaw to a local Ollama instance:
+
+```bash
+openclaw config set provider.ollama.baseUrl http://localhost:11434
+openclaw config set provider.ollama.model llama3:70b
+```
+
+### Multi-Channel Setup
+
+Run OpenClaw with multiple channels simultaneously:
+
+```yaml
+# ~/.openclaw/config.yaml
+channels:
+  telegram:
+    bot_token: ${TELEGRAM_BOT_TOKEN}
+  discord:
+    bot_token: ${DISCORD_BOT_TOKEN}
+  imessage:
+    enabled: true
+```
+
+### Skills
+
+Install community skills:
+
+```bash
+clawhub install web-search
+clawhub install desktop-automation
+clawhub install file-manager
+```
+
+## Troubleshooting
+
+**Environment won't start:**
+```bash
+daytona logs <env-name>
+# Check devcontainer build output
+```
+
+**OpenClaw gateway won't bind:**
+```bash
+# Check port 3000 isn't already in use
+lsof -i :3000
+# Change port in config
+openclaw config set gateway.port 3001
+```
+
+**Claude Code can't see OpenClaw:**
+- Verify the VS Code extension is installed in the Dayton environment (not your local machine)
+- Reload the window after connecting
+
+## Performance Tips
+
+1. **Use prebuilds**: Configure Daytona to prebuild your environment so new branches spin up instantly
+2. **Resource allocation**: Give the Daytona environment at least 2 CPU cores and 4 GB RAM for smooth agent operation
+3. **Persistent storage**: Mount a volume for OpenClaw's memory files so they survive environment restarts
+
+## Comparison: Manual vs. Daytona Setup
+
+| Step | Manual | Daytona |
+|------|--------|---------|
+| Install dependencies | 30–60 min | 0 min (pre-configured) |
+| Configure IDE | 15–20 min | 0 min (pre-configured) |
+| Onboard new developer | 1–2 hours | `daytona create` + 5 min |
+| Environment reproduction | Days of debugging | Identical every time |
+| CI/CD integration | Complex | Built-in |
+
+## Next Steps
+
+- [Daytona documentation](https://daytona.io/docs) — advanced provider configuration
+- [OpenClaw documentation](https://docs.openclaw.ai) — skills, channels, and agent hosting
+- [ClawHub](https://clawhub.ai) — browse community skills
+
+## Conclusion
+
+Daytona + OpenClaw removes the biggest friction point in AI agent setup: environment configuration. Spend your time building skills and automations, not fighting dependency conflicts.
+
+Two minutes. That's all it takes.