This repository automatically builds and publishes Docker images for Archon to GitHub Container Registry. The images are built from source and stay up-to-date with the latest releases.
| Component | Description | Image |
|---|---|---|
| Server | FastAPI backend server | ghcr.io/yourusername/archon-server |
| MCP | Model Context Protocol interface | ghcr.io/yourusername/archon-mcp |
| Agents | AI agent hosting service | ghcr.io/yourusername/archon-agents |
| Frontend | React + Vite UI | ghcr.io/yourusername/archon-frontend |
Replace
yourusernamewith your GitHub username
1. Check for Updates (check-updates.yml)
- Daily monitoring at 6 AM UTC for new commits
- Smart detection - only triggers builds for new commits
- Manual forcing - option to force check/build
- Automatic triggering of build workflow when needed
2. Build and Push Images (build-images.yml)
- Production builds with security features
- Multi-platform support (linux/amd64, linux/arm64)
- Triggered automatically by check workflow or manually
Force a check for updates:
- Go to Actions → Check for Archon Updates
- Click Run workflow
- Check "Force check" if needed
- Click Run workflow
Build specific ref directly:
- Go to Actions → Build and Push Archon Images
- Click Run workflow
- Enter the desired Archon ref (commit, branch, or tag)
- Click Run workflow
Each image is tagged with:
latest- Latest built version{commit-sha}- Specific Archon commit{date}- Build date (YYYY-MM-DD){date}-{short-sha}- Date with commit identifier
Example:
# Pull latest
docker pull ghcr.io/yourusername/archon-server:latest
# Pull specific commit
docker pull ghcr.io/yourusername/archon-server:abc1234567890def...
# Pull by date
docker pull ghcr.io/yourusername/archon-server:2024-01-15
# Pull by date with commit
docker pull ghcr.io/yourusername/archon-server:2024-01-15-abc1234For Kubernetes deployments: See KUBERNETES.md for comprehensive deployment guide addressing common issues like "Blocked request" errors, health check failures, and port configuration problems.
version: '3.8'
services:
archon-server:
image: ghcr.io/yourusername/archon-server:latest
ports:
- "8181:8181"
environment:
- DATABASE_URL=your_database_url
archon-mcp:
image: ghcr.io/yourusername/archon-mcp:latest
ports:
- "8051:8051"
environment:
- ARCHON_MCP_PORT=8051
- ARCHON_SERVER_PORT=8181
- ARCHON_AGENTS_PORT=8052
- VITE_PORT=5173
archon-agents:
image: ghcr.io/yourusername/archon-agents:latest
ports:
- "8052:8052"
environment:
- ARCHON_AGENTS_PORT=8052
archon-frontend:
image: ghcr.io/yourusername/archon-frontend:latest
ports:
- "5173:5173" # FIXED: Use 5173:5173, not 3737:5173# Run server
docker run -p 8181:8181 ghcr.io/yourusername/archon-server:latest
# Run MCP service
docker run -p 8051:8051 -e ARCHON_MCP_PORT=8051 -e ARCHON_SERVER_PORT=8181 -e ARCHON_AGENTS_PORT=8052 -e VITE_PORT=5173 ghcr.io/yourusername/archon-mcp:latest
# Run agents service
docker run -p 8052:8052 -e ARCHON_AGENTS_PORT=8052 ghcr.io/yourusername/archon-agents:latest
# Run frontend - FIXED: Use consistent port mapping
docker run -p 5173:5173 ghcr.io/yourusername/archon-frontend:latestFork this repository to your GitHub account.
Ensure GitHub Actions are enabled in your repository settings.
The workflow uses GitHub Container Registry (GHCR) by default. No additional setup is required - it uses the built-in GITHUB_TOKEN.
Manually trigger the workflow to build your first set of images:
- Go to Actions tab
- Select Build and Push Archon Images
- Click Run workflow
- Leave the ref as
main(or specify a version) - Click Run workflow
- Upstream:
coleam00/Archon - Method: Builds from source (not copying existing images)
- Components: All 4 microservices
The build workflow processes all components in parallel with enhanced Kubernetes support:
- Server: Python FastAPI backend with health endpoint and proper environment defaults
- MCP: Model Context Protocol service with non-root user security
- Agents: AI agent hosting with health checks and logging configuration
- Frontend: React + Vite application with HOST variable support and health endpoint
- Cleaner logs - check and build operations are separate
- Faster feedback - know immediately if new commits exist
- Flexible triggering - build specific refs without checking
- Better monitoring - separate success/failure tracking
- Resource efficiency - only build when needed
- GitHub Actions cache for faster subsequent builds
- Duplicate detection - check workflow prevents unnecessary builds
- Multi-platform builds cached separately
This repository includes specialized Docker images with Kubernetes-specific improvements:
- HOST Variable Support: Automatically configures Vite allowed hosts for custom domains
- Health Endpoint: Dedicated
/healthendpoint for proper health checks - Port Consistency: Uses port 5173 throughout (no more 3737:5173 confusion)
- Non-root Security: Runs as user 1001 for enhanced security
- Health Endpoints: All services include
/healthendpoints where applicable - Environment Defaults: Proper
HOST=0.0.0.0defaults for Kubernetes - Security Context: Non-root users (UID 1001) for all containers
- Logging: Configurable
LOG_LEVELenvironment variable
- ✅ "Blocked request" errors - Fixed with proper HOST configuration
- ✅ Health check failures - Dedicated health endpoints and correct port mapping
- ✅ Port confusion - Consistent 5173 port usage for frontend
- ✅ Security vulnerabilities - Non-root containers with proper user contexts
See KUBERNETES.md for complete deployment guide and troubleshooting.
- 🔐 Container Signing: All images signed with cosign using keyless signing
- 📋 SBOM Generation: Software Bill of Materials included with each image
- 🛡️ Build Attestation: Cryptographic attestation of build provenance
- 🔍 Vulnerability Scanning: Trivy security scanning with results in GitHub Security tab
- 🏗️ Multi-platform: Native support for AMD64 and ARM64 architectures
Rich OCI-compliant metadata including:
- Source repository and commit information
- Build timestamps and versioning
- License information (MIT)
- Component descriptions and documentation links
- Reproducible builds from source code
- Attestation artifacts stored in registry
- Signed provenance linking images to source commits
- Security scanning results available in GitHub Security tab
Verify image signatures:
# Install cosign
go install github.com/sigstore/cosign/v2/cmd/cosign@latest
# Verify image signature
cosign verify ghcr.io/dapperdivers/archon-server:latest \
--certificate-identity-regexp=".*" \
--certificate-oidc-issuer="https://token.actions.githubusercontent.com"Check SBOM:
# Download and view SBOM
cosign download sbom ghcr.io/dapperdivers/archon-server:latestView attestations:
# Download attestation
cosign download attestation ghcr.io/dapperdivers/archon-server:latestEdit the UPSTREAM_REPO environment variable in .github/workflows/build-images.yml:
env:
REGISTRY: ghcr.io
UPSTREAM_REPO: your-fork/Archon # Change thisModify the cron schedule in the workflow:
schedule:
# Currently: daily at 6 AM UTC
- cron: '0 6 * * *'
# Example: twice daily
- cron: '0 6,18 * * *'Extend the metadata section to add custom tags:
tags: |
type=raw,value=latest
type=raw,value=${{ needs.check-releases.outputs.target_ref }}
type=raw,value={{date 'YYYY-MM-DD'}}
type=raw,value=custom-tag # Add your custom tagsImages are published to GitHub Container Registry (ghcr.io) under your account. They will be:
- Public by default (can be changed in package settings)
- Linked to this repository
- Versioned with multiple tags
- Keyless signing with GitHub OIDC (no manual keys required)
- SLSA Level 3 build provenance attestation
- Vulnerability scanning integrated into CI/CD
- Multi-platform builds ensure compatibility
- Reproducible builds from source code
- Supply chain security with SBOM and attestations
- GitHub Security tab integration for vulnerability reports
Check the Actions tab to monitor:
- Build status and logs
- Failed builds and errors
- Build duration and cache performance
- Image publishing status
- Security scan results and attestation generation
- Security tab shows vulnerability scan results
- Packages tab shows published images with signatures
- Attestations are available for each image version
For testing containers locally before deployment, see the tests/ directory which includes:
test-containers.sh- Script to build and test all containers locallydocker-compose.test.yml- Docker Compose configuration for testingREADME.md- Testing documentation and usage instructions
The testing setup replicates the same build process used in CI/CD.
This repository is focused solely on Docker image automation. For Archon-related issues or contributions, please visit the upstream repository.