Aigentron

Clone the repository

git clone https://github.com/Arttron/aigentron.git
cd aigentron

Prerequisites

• Docker Desktop (or Docker Engine + Compose v2)
• Ollama installed and running on the host (not in a container) — powers the 'routine' model tier. The orchestrator container reaches it at http://host.docker.internal:11434
• ANTHROPIC_API_KEY — required for the 'complex' model tier (cloud Claude)
• A tool-capable model in your host Ollama (default qwen3-coder:30b, set via ROUTINE_MODEL in .env)
• A one-time external Docker network so agents can reach project databases by service name: docker network create lds-agents
• Avoid qwen2.5-coder — it doesn't emit structured tool calls and breaks the agent SDK. qwen3-coder:30b is the supported default.

1. Configure environment

Fill in at minimum ANTHROPIC_API_KEY and ROUTINE_MODEL (defaults are correct for local use)

cp .env.example .env

2. Pull the routine-tier model

Pull into your host Ollama (not into a container). Set ROUTINE_MODEL in .env if you use a different model.

ollama pull qwen3-coder:30b

3. Create external network

One-time setup — agents connect to project databases via service names on this network

docker network create lds-agents

4. Build and start

Equivalent to docker compose up -d --build. Brings up Postgres, Redis, LiteLLM (Anthropic-compatible proxy in front of Ollama), the orchestrator (:3001), and the dashboard (:3000), all bind-mounted for hot-reload. First boot is slower — the orchestrator compiles packages/shared and packages/agent-runner before its healthcheck passes.

make up

5. Open dashboard

Dashboard → http://localhost:3000 • Orchestrator API → http://localhost:3001 Optional: docker compose --profile mcp up -d for Playwright MCP and a dev server for frontend/design agents.