Overview
All PicoClaw configuration options can be overridden using environment variables. This is useful for:- Containerized deployments (Docker, Kubernetes)
- CI/CD pipelines
- Temporary overrides without editing config files
- Secret management (API keys, tokens)
Environment Variable Format
Environment variables follow this pattern:PICOCLAW_AGENTS_DEFAULTS_MODEL_NAMEPICOCLAW_TOOLS_WEB_BRAVE_API_KEYPICOCLAW_HEARTBEAT_ENABLED
Core Configuration
PICOCLAW_CONFIG
Purpose: Override config file location~/.picoclaw/config.json.
PICOCLAW_HOME
Purpose: Override root directory for PicoClaw data- Default workspace to
/opt/picoclaw/workspace - Default config to
/opt/picoclaw/config.json(if not usingPICOCLAW_CONFIG)
Agent Configuration
Workspace and Security
Model Configuration
Agent Behavior
Provider Configuration (Legacy)
OpenAI
Anthropic
Zhipu (智谱)
OpenRouter
Groq
Gemini
Other Providers
Tools Configuration
Web Search - Brave
Web Search - Tavily
Web Search - DuckDuckGo
Web Tools General
Cron Tools
Exec Tools
MCP
Skills Registry
Media Cleanup
Gateway Configuration
Heartbeat Configuration
Devices Configuration
Channel Configuration
Telegram
Discord
DingTalk
LINE
WeCom
Usage Examples
Docker Compose
Kubernetes Secret
Shell Script
systemd Service
Best Practices
- Use secrets management for API keys in production
- Never commit
.envfiles with real credentials - Use read-only environment variables in containers
- Override selectively - only set what you need to change
- Document environment variable usage in deployment docs
Precedence
Configuration resolution order:- Environment Variables (highest priority)
- Config File (
~/.picoclaw/config.json) - Default Values (lowest priority)