Codewhale CLI
An open terminal coding agent in Rust: reads the project, edits files and runs commands on any model, including a local one through Ollama
- 41,140 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Edits project files and runs commands in the terminal
- Installing by an internet script runs someone else's code
- Requests go to the chosen model provider
Install
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/codewhaleDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Edits project files and runs commands in the terminal
- Installing by an internet script runs someone else's code
- Requests go to the chosen model provider
This is third-party code. Review the repository files before installing.
What it does
Codewhale is an open agent for working with code in the terminal. It reads the project, edits files and runs commands, and you choose the model: hosted at a provider or local. If Ollama is running, the agent switches to it on its own. The `/provider` and `/model` commands change the provider and model inside a session. The project is written in Rust, version 0.10.0 from 22.09.2026. Installation: `curl -fsSL https://codewhale.net/install.sh | sh`, `npm install -g codewhale`, `winget install HunterBown.CodeWhale` or `cargo install codewhale-cli --locked`, updated with `codewhale update`.
Who it is for. For developers who need a terminal agent with no tie to one provider: you can use different models and work fully locally.
Good fit when
- You need a coding agent that works with a model of your choice rather than one
- You need cloud-free work on a local model through Ollama
- You need a single agent for macOS, Linux and Windows installed through winget or npm
Not a fit when
- You need the catalog's skills and MCP servers for a specific agent: check support in the Codewhale documentation
- You are not ready to let an agent edit files and run commands in a project
Example request
Find the function in the project that fails on an empty list, fix it and run the testsLimitations
The agent edits project files and runs commands in the terminal, so work on a separate branch and read the proposed actions. Installing through `curl ... | sh` runs a script from the internet: read it before running. Quality depends on the chosen model. We did not check skill and MCP support in the agent's documentation. The repository belongs to the codewhale-hq organization and used to live in the hmbown profile.
How to disable. Remove the installed package (npm uninstall -g codewhale, winget uninstall or cargo uninstall) and revoke the provider keys you entered.
FAQ
Which models does Codewhale support?
Any: hosted at a provider and local. If Ollama is running, the agent switches to it on its own.
How do I install it?
With the curl script, npm install -g codewhale, winget install HunterBown.CodeWhale or cargo install codewhale-cli --locked.
How do I change the model during work?
With the /provider and /model commands inside the session.
How do I update it?
With the codewhale update command.
Discussion
No questions or reviews yet.
Related
CLI toolsA lightweight Rust coding agent for the terminal, inspired by pi and OpenCode and built for a small memory footprint
An open coding agent for the terminal, browser and desktop that works with models from OpenAI and Claude to Ollama
A self-hosted runtime for coding agents: each crew in its own Linux container, wrapped in a control plane with roles and audit
An open-source terminal coding agent from the Qwen team that works with Qwen, OpenAI, Anthropic, Gemini and local models
A Go coding agent for open models: it turns a chat into tasks on separate branches, checks and merges the result and runs all projects in one window
A terminal coding agent built on Codex, tuned for low-cost open models like Kimi K3, DeepSeek and Qwen
README in short
The README describes Codewhale as an open terminal coding agent in Rust, with a choice of model and provider, local models through Ollama and installation through a script, npm, winget or cargo. MIT license.