CodeWhale
Open-source, community-driven agent harness
CodeWhale is an open-source terminal-based AI coding agent written in Rust that brings your own model for autonomous code editing, command execution, and task completion. Originally built for DeepSeek, it now supports 30+ providers including Claude, GPT, Kimi, GLM, and local models through vLLM, SGLang, or Ollama. The agent's unique architecture lets you assign different models to different roles within a single task—a cheap model can orchestrate while an expensive reasoning model executes—and you can switch models mid-task or customize roles and permissions through editable configuration files. It runs interactively through a TUI, headlessly via `codewhale exec`, or through a local web interface.
Key Features
Use Cases
- 01Autonomous debugging and test fixing by running `codewhale exec "fix the failing test"` in CI pipelines
- 02Multi-model code review workflows where a fast model triages changes and a reasoning model performs deep analysis
- 03Cost-optimized development by routing simple tasks to cheaper models and complex reasoning to premium tiers
- 04Offline development with local models through Ollama or vLLM when API access is unavailable
- 05Collaborative team coding with shared role configurations and constitutions stored in project repositories
- 06Interactive code refactoring with workspace snapshots and `/undo` or `/restore` for safe experimentation
Related Agents
View morehermes-agent
The agent that grows with you
agency-agents
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
openinterpreter
A coding agent for open models like Kimi K3
cline
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
CodeWhale — FAQ
What is CodeWhale?+
CodeWhale is an open-source terminal coding agent written in Rust that autonomously reads code, edits files, runs commands, and checks its work using AI models you provide. It supports over 30 providers and local models, with the unique ability to assign different models to different roles within a single task.
How do I install CodeWhale?+
Install globally via npm with `npm install -g codewhale`, or use Cargo, Docker, Nix, Scoop, or prebuilt binaries as detailed in the project's INSTALL.md documentation. After installation, authenticate with `codewhale auth set --provider <name>` or export provider API keys as environment variables.
Which AI models and providers does CodeWhale support?+
CodeWhale supports 30+ providers including DeepSeek, Anthropic Claude, OpenAI GPT, Kimi, GLM, xAI Grok, OrcaRouter, and local setups via vLLM, SGLang, or Ollama. You can mix providers and models within a single task by assigning different models to different roles.
Do I need API keys to use CodeWhale?+
You need API keys for hosted providers like DeepSeek, Anthropic, or OpenAI, which you configure with `codewhale auth set` or environment variables. For local models through vLLM, SGLang, or Ollama, no API keys are required.
Is CodeWhale free to use?+
CodeWhale itself is free and open-source under the MIT license. However, you'll incur costs from whichever AI model providers you use unless you run local models through Ollama or similar self-hosted solutions.
Can I use CodeWhale in CI/CD pipelines?+
Yes, CodeWhale offers a headless `codewhale exec` mode designed for scripting and continuous integration environments. It executes tasks non-interactively and returns results suitable for automated workflows.
How do I install CodeWhale?+
Open the source repository on GitHub and follow its README. CodeWhale is a agent — MCP Agents Market links you directly to the official repo.
Is CodeWhale free?+
CodeWhale is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.