AgentIndex icon
AgentIndex
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Submit Tool
ToolsMCP ServersHOTCategoriesTrendingNewBest ListsCompareGuides
Home/
Code Assistant/
Codewhale
Codewhale logo

Codewhale

Active·★ 41.1k·MIT·Updated 2026-10-22
★ Most Popular

Codewhale is an open-source coding agent for the terminal, written in Rust. Give it a task and it reads the repository, edits files and runs commands until the work is done or it needs input. It ships more than 40 providers including Anthropic, OpenAI, DeepSeek, Google, Mistral, Moonshot, OpenRouter and xAI, accepts any OpenAI-compatible endpoint, and uses local models through Ollama, vLLM or SGLang; DeepSeek's flash model is the default.

Control comes from three modes and three permission levels. Plan only reads, Work edits and runs commands, and Operate keeps going until a goal is verified. Ask applies file edits but asks before non-read-only shell commands, Auto-Review lets rules and a reviewer model approve routine calls, and Full Access turns the sandbox off. /receipts lists every file changed, command run, web and MCP call, sub-agent and approval; /undo reverts the last request and /restore rolls the workspace back to a snapshot.

It also supports sub-agents, fleets and repository workflows for longer jobs, reads AGENTS.md automatically, and runs headless in CI with codewhale exec. Note the defaults: Ask still applies file edits without asking, and exec --auto means Full Access unless you add --sandbox workspace-write, so commit or stash first.

This write-up is based on the project's README and public repository data. We have not hand-tested the tool.

#coding-agent#terminal#rust#cli#deepseek#ollama#local-first
$ Install
$ curl -fsSL https://codewhale.net/install.sh | sh
↗ Visit site★ GitHub
01

Features

0140+ built-in providers, any OpenAI-compatible endpoint, and local models via Ollama, vLLM or SGLang
02Plan, Work and Operate modes with Ask, Auto-Review and Full Access permissions
03/receipts audit trail of files, commands, web and MCP calls and approvals
04/undo and /restore to roll back a request or the whole workspace
05codewhale exec for headless runs in scripts and CI, with an optional workspace sandbox
02

Why choose it

+Works with 40+ providers and local models, so you can run it on cheap or self-hosted models.
+/receipts gives a full audit trail of files, commands, web and MCP calls and approvals.
+/undo and /restore make it easy to back out of a bad run.
03

Trade-offs

!Default Ask mode applies file edits without asking, and exec --auto runs with Full Access unless sandboxed.
!Fast release cadence: some features differ between versions, such as /undo needing /trust on in 0.10.1.
!Results depend heavily on the model you choose, and the default is a budget model.
04

Maintenance check

ActiveChecked: 2026-10-23
Last commit
2026-10-22
Commits, last 90 days
100+
Releases, last 90 days
15
Latest release
v0.10.1 (2026-10-08)
Open issues
209
Recent issues with at least one comment
5/17
Commits per week, last 12 weeks
Commits per week, last 12 weeks: 519, 98, 417, 160, 241, 222, 293, 788, 738, 231, 267, 204

Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.

Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.

Most discussed open issues

  • #5316 EPIC-005: CodeWhale TUI Crate Decomposition (Umbrella)33 comments · 2026-08-10
  • #6011 feat(tui): usage & tool diagnostics — token accounting (per-component, per-model + cache hit rate, per-tool sinks, compaction cost) and tool10 comments · 2026-09-08
  • #6978 Turn snapshots fail with "git commit-tree failed: fatal: <sha> is not a valid object" since 2026-10-12 (0.10.0 after rollback from 0.10.1); 9 comments · 2026-10-14
  • #6184 Engine silently freezes mid-run: user messages are persisted but never answered; no error, no log line, no crash entry9 comments · 2026-09-15
  • #6094 Start here: what the next patch release covers, and how to help9 comments · 2026-09-11

Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.

05

Compatibility

macOS / Linux / Windows
CLI (x64, arm64)
Verified via docs
Ollama
Local models
Verified via docs
MCP
MCP calls
Verified via docs
06

Quick start

1
$ curl -fsSL https://codewhale.net/install.sh | sh
07

Use cases

↳A terminal coding agent on cheaper or local models
↳Fixing failing tests or refactoring with an audit trail of what changed
↳Running agent tasks in CI with codewhale exec
08

Alternatives

⚡Top VS:
VSGemini CLI★107.5k→VSkimun★62→VSclaude-code-guide★4.7k→VSGraft★10.7k→
Gemini CLI logo
Gemini CLI★ 107.5k
An open-source AI agent that brings the power of Gemini directly into your terminal. Supports native MCP.
VSCompare→
kimun logo
kimun★ 62
Simple note taking, powerful search, AI ready
VSCompare→
claude-code-guide logo
claude-code-guide★ 4.7k
Master Claude Code with this Guide! Includes: Setup, SKILL.md files, Agents, Commands, workflows and tricks making Claude's potential skyrocket!
VSCompare→
Graft logo
Graft★ 10.7k
Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.
VSCompare→
Context7 logo
Context7★ 63.5k
MCP Server that provides up-to-date code documentation for LLMs and AI code editors.
VSCompare→
Open Interpreter logo
Open Interpreter★ 68.6k
A coding agent for the terminal, optimized for low-cost and open models. Forked from OpenAI's Codex, it can emulate other agent harnesses and runs commands in a native sandbox.
VSCompare→
learn-claude-code logo
learn-claude-code★ 79.0k
Bash is all You need - Write a nano Claude Code 0 - 1
VSCompare→
awesome-cursorrules logo
awesome-cursorrules★ 41.0k
📄 Configuration files that enhance Cursor AI editor experience with custom rules and behaviors
VSCompare→
See all alternatives →

Comments

Log in to leave a comment

No comments yet. Be the first!

On this page
01Features02Why choose it03Trade-offs04Maintenance check05Compatibility06Quick start07Use cases08Alternatives
codewhale-hq /
Codewhale
↗

A Rust terminal coding agent for any model, with receipts of everything it did and one-step undo

41.1kMIT
Last commit1d ago
StatusActive
CategoryCode Assistant
Links
Documentation↗Discussion↗Issues↗Releases↗
Cloud Infrastructure
Ad

Deploy on DigitalOcean

Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.

Claim $200 Free Credit
© 2026 AgentIndex.app|Built by a 10-year iOS Developer.
Best by use caseAboutContactAdvertisePrivacy PolicyTerms of UseQYSGitHubBuy me a coffee ☕

Browse by Category

MCP ServersCode AssistantWorkflow AutomationRAG / Knowledge BaseMulti-AgentBrowser AutomationLLM InfraDev ToolingObservability

Not affiliated with Anthropic, OpenAI or Microsoft.