Codewhale documentation
ReleaseLatest release v0.10.2 · Oct 24, 2026These pages describe v0.10.2, the published release.Changelog →

Install and use Codewhale

Find installation steps, model setup, tool configuration and troubleshooting. The guides include commands and examples you can run.

By task

23

Start from what you are trying to do.

Inspect the outbound request

Read a redacted request manifest without sending or changing session state.

Read language-server diagnostics

Return LSP errors after edits and configure custom servers by file extension.

Inspect execution policy rules

Read rule sources, matchers and scope; understand allow, ask and deny precedence.

Manage reusable skills

Find SKILL.md folders, import into owned roots and control invocation.

Move from Claude Code

Keep compatible instructions and skills, and review a settings import plan.

Install Codewhale

One command on macOS or Linux; npm, Cargo, release binaries, and the Homebrew tap on Linux also work.

Start your first task

Open Codewhale in a project, look around in Plan, then let it edit in Work.

Connect a provider

Save a key for DeepSeek, OpenAI, Anthropic, OpenRouter, or another provider — or run a local model with no key.

Choose or switch models

Supported providers, switching mid-session, and local runners.

Decide what runs without asking

Plan, Work, or Operate; Ask, Auto-Review, or Full Access; and how to answer a prompt.

Review what changed

See every edit, roll files back to an earlier turn, and get a code review before you push.

See what a session did

Read a session receipt of commands, file changes, tools and approval decisions.

Watch each session at a glance

Turn on pet mode in each terminal to see which session is working, waiting for you, done, or failed.

Run a workflow

Write a repeatable Workflow, run it as a Lane, and watch or stop it from any terminal.

Run agents in parallel

Roles, separate worktrees for parallel edits, and how to watch sub-agents.

Connect tools with MCP

Add a local or remote MCP server, sign in with OAuth, or serve Codewhale over MCP.

Try code mode

Let the model compose several tool calls — MCP and plugin tools included — in one short program, under the same approvals as direct calls. Experimental, on by default.

Run commands on events

Block a risky command, add context, or get notified when Codewhale waits for you.

Limit what commands can touch

See which OS sandbox your platform has and turn on bubblewrap on Linux.

Automate with the Runtime API

Run one-shot jobs in CI, or drive threads and approvals over the local HTTP API.

Open the browser client

Work in a local browser tab, or continue a session from the web app with /rc.

Fix a problem

Diagnose in one command, then fix install, key, network, and stuck-turn problems.

See what leaves your machine

What a provider receives, what usage counting sends, and how to turn it off.

By topic

25

Get started

6

Work with Codewhale

8

Extend and automate

6

Get help

2

Reference

3

Web guides open on codewhale.net. Source docs open the full reference in the GitHub repository.