Verified project record

codewhale-hq/Codewhale

CodeWhale is an open-source terminal coding agent that reads projects, edits files, runs commands, and checks its own work using a hosted or local model of the user's choice. It offers over 40 built-in provider routes plus local runtimes such as Ollama, vLLM, and SGLang.

Project overview

It addresses provider lock-in directly: the same agent works with over 40 provider routes, any OpenAI-compatible endpoint, or local runtimes, so the model choice stays with the user.

Project type: AI Agent · AI Coding
Use cases: Coding & Development · Automation
Deployment: Refer to project documentation
License: MIT

Best for

Developers who want an open-source terminal coding agent and the ability to choose among hosted providers or local model runtimes.
Users who prefer natural-language task input without writing code to drive the agent.

Key capabilities

Reads the project, edits files, runs commands, and checks its own work in the terminal with a hosted or local model of the user's choice, via over 40 built-in provider routes and local runtimes.

Limitations and risks

Usage-counts telemetry is on by default; it can be disabled, but the exact telemetry scope is not fully documented in the available facts. Teams with strict data policies should review and disable it during setup.
A hosted provider key or a local model runtime must be configured before use. Costs of hosted providers and GPU or hardware requirements for local models are not documented in the available facts.

Getting started

Install with the curl installer, which downloads checksum-verified binaries: curl -fsSL https://codewhale.net/install.sh | sh. Then run codewhale in the project folder, use /provider to connect a model, and give a task. Setup is rated easy because of the single installer and natural-language task input.

Alternatives and comparisons

A self-hosted developer control center for running coding agents across local, Docker, VM, and cloud backends, suited when multi-environment agent orchestration is needed rather than a single terminal agent.
A unified multi-provider LLM API with an interactive terminal coding agent CLI, relevant when a multi-provider API layer plus agent runtime is wanted.

Project comparisons

Evidence and sources

README: Codewhale reads your project, edits files, runs commands, and checks its own work — in your terminal, with a hosted or local model you choose.
Release: v0.10.0
README: curl -fsSL https://codewhale.net/install.sh | sh
README: Run `/provider` (or press `F3`) to add a hosted key or pick a local runtime.
README: `codewhale exec "…"` | One headless turn from a script or CI, streaming JSON

Source updated at: 2026-10-05T18:16:37Z

Open original GitHub project