Basic Configuration

FinClaw runtime configuration lives in the .env file in the deploy directory (generated by init-env.sh from .env.example). On first deployment the key entries below ship in the offline package and work out of the box.

Variable	Description	Source
DEPLOYMENT_LICENSE_KEY	License key	Auto-generated by init-env.sh on first deployment (the offline package's .env.example ships with the Finogeeks-issued value; no manual setup); for renewal see License Update
DEPLOYMENT_LICENSE_CIPHER	License cipher	Same as above

LLM Model (Admin Console; env Is Bootstrap Only)

Model configuration (provider, model name, API key, base URL) is primarily managed in the admin console (Admin), persisted in the database, with multi-model support. The .env variables LLM_PROVIDER / LLM_MODEL / LLM_API_KEY / LLM_BASE_URL serve only as first-startup bootstrap and fallback (defaults LLM_PROVIDER=deepseek, LLM_MODEL=deepseek-v4-flash); no manual configuration needed.

See Business Configuration.

Custom External Port

FinClaw exposes the host port 26100 through the API Gateway container by default (26100:26100 on the api-gateway service in docker-compose.yml). If that port is occupied, or you need a unified ops port plan, remap it as follows (example: change to 28100).

Using docker-compose.override.yml

Does not modify the original docker-compose.yml, so it survives upgrades.

cd /data/chatkit-offline-release-<version>-linux-amd64/deploy

cat > docker-compose.override.yml <<'EOF'
services:
  api-gateway:
    ports:
      - "28100:26100"   # left = host port, right = container port (do not change)
EOF

# Restart to apply
./scripts/fleet.sh down
./scripts/fleet.sh up -d

Next-Step Configuration Guide

To configure	Go to
LLM multi-model, channels (Feishu/WeCom/DingTalk), persona templates, skill service, administrator & session, storage & upload	Business Configuration
Concurrency quotas, agent loop, execution posture & FinSAFE/Hermes, auth tokens, claw-fleet-manager, observability	Operations Configuration