Configuration

Core Environment Variables

• OPENCLAW_IMAGE: OpenClaw image to run
• OPENCLAW_PODMAN_PUBLISH_HOST: host publish address
• OPENCLAW_OLLAMA_BASE_URL: Ollama native API URL
• OPENCLAW_OLLAMA_MODEL: default Ollama model id
• OPENCLAW_SCALE_INSTANCE_ROOT: root directory for generated instances
• OPENCLAW_SCALE_GATEWAY_PORT_START: first gateway port
• OPENCLAW_SCALE_BRIDGE_PORT_START: first bridge port
• OPENCLAW_SCALE_PORT_STEP: per-instance port increment

Agent-Team Tuning

These env vars control the communication lab:

• OPENCLAW_MATTERMOST_ENABLED
• OPENCLAW_MATTERMOST_CHANNEL_NAME
• OPENCLAW_MATTERMOST_AUTONOMY_ENABLED
• OPENCLAW_MATTERMOST_AUTONOMY_INTERVAL
• OPENCLAW_MATTERMOST_AUTONOMY_INTERVAL_INSTANCE_00N
• OPENCLAW_MATTERMOST_AUTONOMY_MODEL

Most team behavior still lives in the workspace scaffolds rather than env vars. Use .env for infrastructure defaults, then tune SOUL.md, IDENTITY.md, USER.md, and HEARTBEAT.md inside each generated workspace.

Important Note For Windows + WSL Podman

The working Ollama endpoint observed during validation was:

http://172.27.208.1:11434

When .env still uses the default host.containers.internal value, generated runtime config now auto-resolves it to the current Podman-machine gateway on Windows when that host alias is not usable. .\scripts\doctor.ps1 also fails fast if the effective Ollama /api/tags endpoint is unreachable.

If you want to pin a different Ollama endpoint yourself, replace OPENCLAW_OLLAMA_BASE_URL accordingly.

OPENCLAW_MATTERMOST_AUTONOMY_MODEL is intentionally blank in .env.example, so mattermost lounge enable inherits the current primary model unless you set an explicit override.

Trust Model

This project targets same-trust operator workflows. OpenClaw is configured in a full-access-in-container mode and depends on the Podman pod boundary rather than OpenClaw's internal sandbox for the main isolation layer.