Self-Hosted

Availability

Self-hosted CodeDig is suited for organisations that:

• Cannot send source code to external services (compliance, ITAR, etc.)
• Require an air-gapped or private-network deployment
• Need to integrate with an on-premises GitHub Enterprise or GitLab instance
• Have contractual data residency requirements

The self-hosted deployment is functionally identical to the SaaS offering. All analysis runs locally — no code leaves your network.

Deployment artifacts

After your Enterprise agreement is in place, we provide:

• Container images — published to a private registry; pull with the credentials supplied by your account team
• Helm chart — at helm/archaeologist-operator/ in the release bundle (values.yaml, values-prod.yaml, values-staging.yaml)
• Docker Compose reference — docker/docker-compose.prod.yml for non-Kubernetes deployments
• License key — required at startup; loaded via the CODEDIG_LICENSE_KEY environment variable

Requirements

Docker Compose

For smaller teams or evaluation deployments, Docker Compose is the fastest path to a running instance. A production-ready reference is included in the release bundle:

# Reference: docker/docker-compose.prod.yml
services:
  api:
    image: ghcr.io/sassy-solutions/codedig-api:latest
    environment:
      DATABASE_URL: postgres://user:pass@db:5432/codedig
      REDIS_URL: redis://redis:6379
    ports:
      - "4000:4000"
    depends_on: [db, redis]

  db:
    image: postgres:16
    environment:
      POSTGRES_DB: codedig
      POSTGRES_USER: user
      POSTGRES_PASSWORD: pass

  redis:
    image: redis:7-alpine

See docker/docker-compose.prod.yml in the release bundle for the complete definition including volume mounts, health checks, and resource limits.

Kubernetes / Helm

For production at scale, use the bundled Helm chart. Override values as needed for your environment:

# Reference: helm/archaeologist-operator/
helm install codedig ./helm/archaeologist-operator \
  -f helm/archaeologist-operator/values-prod.yaml \
  --set api.env.DATABASE_URL="postgres://..." \
  --set api.env.REDIS_URL="redis://..." \
  --namespace codedig --create-namespace

The chart includes a Kubernetes Operator (crates/operator/) for managing CodeDig resources as Kubernetes custom resources, enabling GitOps-style configuration management.

Air-gapped environments

CodeDig's core analysis pipeline runs entirely locally and requires no outbound internet access for parsing, security scanning, or PR gate analysis. However, two features require outbound HTTPS:

• AI-assisted explanations and recommendations — these call your configured LLM provider endpoint. In air-gapped environments, point LLM_BASE_URL at a locally-hosted model (e.g. vLLM or Ollama). Alternatively, AI features can be disabled entirely via the admin console.
• License validation — requires a one-time call to license.codedig.ai at startup. Offline license validation is available for fully air-gapped deployments — contact your account team to enable it.

For more information on AI data handling, see the Privacy Policy → AI Processing section.