Cogy Code
Cogy Code works with any model that has an OpenAI-compatible endpoint. The installer sets up llama.cpp, and MLX on Apple silicon.
The installer configures llama.cpp for the graphics card it finds, and adds MLX and Splash on Apple silicon. Models can be downloaded from Hugging Face inside the app.
Each model has its own settings. Agents can also use the shared models on Cogy Server, so a team can work from one powerful machine.
More in Cogy Code.
An open-source coding agent and browser IDE that runs in Docker, uses local models and connects to Cogy Server.
Agent Sandboxes
Each agent runs in its own Docker Sandbox with its own clone
Permission Modes
Five settings for how much an agent can do without asking
Code Index
An index of the repository for the agent to search
Multi-Agent Jobs
Planner, worker, verifier and reviewer roles on each job
Codex and Claude Code
Run Codex and Claude Code with Cogy Code permissions
Integration Services
Our rollout, model setup and support for Cogy Code
Let us build your future
Bring us a plan or just an idea. We will arrange a discovery meeting with you to go through it.