OpenDevin

OpenDevin is a platform for autonomous software engineers, powered by AI and LLMs. OpenDevin agents collaborate with human developers to write code, fix bugs, and ship features.

https://github.com/OpenDevin/OpenDevin
https://docs.all-hands.dev/modules/usage/intro
https://docs.all-hands.dev/modules/usage/llms/localLLMs

Docker compose file for local setup with separate ollama docker instance. In the current OpenDevin version (0.7), the ollama model needs to be specifically entered into the compose file and/or the interface.

mkdir /opt/opendevin/base
mkdir /opt/opendevin/logs

docker-opendevin.yml

 
name: opendevin
services:
  opendevin:
    image: ghcr.io/opendevin/opendevin:0.7
      #image: ghcr.io/opendevin/opendevin:latest
    container_name: opendevin
    deploy:
      resources:
        reservations:
          devices:
          - driver: nvidia
            device_ids: ['0']
            capabilities: [gpu]
    ports:
      - 3300:3000
    volumes:
      - /opt/opendevin/base:/opt/workspace_base
      - /opt/opendevin/logs:/app/logs
      - /var/run/docker.sock:/var/run/docker.sock
    restart: unless-stopped
    environment:
      - OPENAI_API_KEY=anything
      - CUSTOM_LLM_PROVIDER=ollama
        #- LLM_MODEL=deepseek-coder-v2:latest
      - LLM_MODEL=codestral:latest
      - LLM_EMBEDDING_MODEL="local"
      - LLM_API_KEY=ollama
      - LLM_BASE_URL=http://host.docker.internal:11434
      - LLM_OLLAMA_BASE_URL=http://host.docker.internal:11434
      - TZ=Europe/London
      - SANDBOX_USER_ID=0
      - WORKSPACE_MOUNT_PATH=/opt/opendevin/base/workspace
    extra_hosts:
      host.docker.internal: host-gateway