Quick Start

Prerequisites

• Docker & Docker Compose — For running Condrix services
• Git — For cloning repositories and workspace management

Want to develop Condrix itself?

If you’re contributing to Condrix or running from source, see the Development Commands page for native setup with hot-reload.

Start with Docker Compose

Create a docker-compose.yml file:

services:
  maestro:
    image: ghcr.io/anastawfik/condrix-maestro:latest
    ports:
      - "9200:9200"
    volumes:
      - maestro-data:/data
    environment:
      - CONDRIX_MAESTRO_PORT=9200

  core:
    image: ghcr.io/anastawfik/condrix-core:latest
    ports:
      - "9100:9100"
    volumes:
      - core-data:/data
      - claude-data:/root/.claude
    environment:
      - CONDRIX_CORE_PORT=9100
      - CONDRIX_CORE_HOST=0.0.0.0
      - CONDRIX_MAESTRO_URL=ws://maestro:9200

  web:
    image: ghcr.io/anastawfik/condrix-web:latest
    ports:
      - "5173:80"

volumes:
  maestro-data:
  core-data:
  claude-data:

Start all services:

docker compose up -d

Open http://localhost:5173 in your browser. The Core will be listed in the sidebar — click it to connect.

Configure an AI Provider

Before you can start an AI agent session, the Core needs at least one AI provider connection. Condrix supports Claude, OpenAI, Ollama (local), and any OpenAI-compatible endpoint.

1. Open Settings (gear icon) in the web client
2. Navigate to the AI tab
3. Click Add Connection and choose a provider:
   • Claude (OAuth) — Sign in with your Claude account (recommended)
   • Claude (API Key) — Paste an Anthropic API key
   • OpenAI — Paste an OpenAI API key
   • Local — Connect to Ollama or LM Studio (default: localhost:11434)
   • Custom — Any OpenAI-compatible endpoint with a custom base URL
4. Follow the provider-specific prompts to complete setup
5. The Core is now ready to create agent sessions

For advanced configuration — including fallback chains and per-project profiles — see AI Providers.

Default Ports

Service	Port	Description
Core	9100	Agent runtime daemon (WebSocket)
Maestro	9200	Orchestration service (WebSocket)
Web Client	5173	Browser-based UI (HTTP)
Docs	5174	Documentation site (HTTP)

What’s Next?

• Authentication — AI connections, profiles, and Core auth
• AI Providers — Multi-provider setup with fallback chains
• Architecture Overview — Understand how the pieces fit together
• Environment Variables — Configure ports, hosts, and features