Codewhale is an open-source coding agent for your terminal that automates development tasks by reading projects, editing files, running commands, and validating its work.
The tool solves the problem of repetitive or complex coding work by acting as an autonomous agent that can break down larger tasks into smaller ones and delegate them to multiple agents with different models and roles. It reads your project context, makes edits, executes commands, and reviews its own output to ensure correctness. You choose whether to use a hosted model or run one locally, and the agent operates within permission modes you set—from planning-only exploration to full file and command execution.
Codewhale suits developers who want to automate coding tasks from the terminal without leaving their workflow. It works well for projects where you can describe concrete tasks and want the agent to handle implementation details. The tool is local-first, meaning your code stays on your machine by default, and it supports multiple models and providers so you are not locked into a single service. It integrates with Ollama for local model inference and can work with hosted providers via API keys.
The project maintains active engagement with its community through an open issue and pull request process. Development includes regular updates to the changelog documenting both released and unreleased changes. The tool supports installation across multiple platforms and package managers, with ongoing migration support for users switching between installation methods. The codebase is written in Rust and includes a terminal user interface that receives iterative refinement across development builds.
