Terminal coding agent
- Price
- $17/mo billed yearly · $20 monthly
- License
- Proprietary
- Stars
- 156k
- Shipped
- 1d ago
Terminal coding agent
In the maker’s words: mu (μ): most of what a coding agent decides isn't code. mu hands 40 kinds of yes/no calls to a small model: fewer wasted tokens, fewer approvals on autopilot, fewer runs that stop halfway. Built on pi.
api.github.com ↗
Not yet verified. An open-source license does not tell us what a hosted plan costs, so we do not infer one.
Counts cover stable published GitHub releases. Charts show days with at least one stable product release, so multiple releases in one day count once. Prereleases, declared nightlies and timestamped or commit-artifact builds are counted separately. Drafts and bare tags are excluded. Monorepo counts cover the verified product stream, excluding SDKs, infrastructure and sibling packages. All windows end at the recorded release snapshot. Release methodology. Collected 18 Nov 2026. Release evidence
Release days per week, last 26 weeks
A score needs at least one Hacker News mention in the last 90 days. mu has none yet, so it has no score and no rank. Upvotes still show their real count.
609 GitHub stars, read 18 Nov 2026.
Fewer than 10 votes. Votes can only raise a score: up to 20% of it reflects upvotes, graded on a curve against the most-voted tool. The square-root curve has a floor of 100 effective votes. Votes decay with a 90-day half-life. Maker self-votes do not count. Counts below 10 are hidden for display only.
We are gathering what developers say about mu on X, Reddit and Hacker News. Classifying it is not live yet, so there is no sentiment score and sentiment is not part of the Index score. When it is, this section shows the positive, neutral and negative split with a margin.
A sample of threads found during research: issues and discussions, not a measure of satisfaction.
mu (μ): most of what a coding agent decides isn't code. mu hands 40 kinds of yes/no calls to a small model: fewer wasted tokens, fewer approvals on autopilot, fewer runs that stop halfway. Built on pi.api.github.com ↗ · read 18 Nov 2026
Most of what a coding agent decides isn't code. It's yes/no calls, and mu hands 41 kinds of them to a small model. Built on pi .github.com ↗ · read 18 Nov 2026
Terminal-Bench 2.1, same model, same tasks: mu passed 68.2% of attempts, pi 58.7% (+9.5 points, 95% interval +3.8 to +15.2). How it was measuredgithub.com ↗ · read 18 Nov 2026
<p align="center"><code>npm i -g mu-agent && mu setup && mu</code> · Node.js 22.19 or newer · or the <a href="https://github.com/qybaihe/mu/releases/latest">desktop app</a> for macOS, Windows and Linux</p>github.com ↗ · read 18 Nov 2026
<img src="docs/readme/run-light.gif" alt="A real run in the mu desktop app, sped up, in mu’s own repository. The message: mu “- fix the parser” fails with “Unknown option”; a task that starts with “- ” is a message, not an option; the CLI is in packages/coding…github.com ↗ · read 18 Nov 2026
pi's own commands (`/model`, `/thinking`, `/login`, `/resume`, `/tree`, `/fork`, `/compact`, `/export` and the rest) work as they do in pi, except that `/login` offers no Claude subscription sign-in: Anthropic doesn't allow other apps to offer one, so use an A…github.com ↗ · read 18 Nov 2026
OpenAI's decision model, gpt-6-luna, answers through OpenAI's Decisions API with an OpenAI API key: MU_JUDGE_OPENAI_KEY, or the OpenAI key mu already has. Pick it with mu setup --judge openai or under Settings → Judges (details).github.com ↗ · read 18 Nov 2026
A bar along the bottom keeps the kernel in view: which judge answers and where it runs, how many decision points are active and, in a conversation, the verdicts so far, the latest judge latency and the context in use. The permission mode and the goal sit in th…github.com ↗ · read 18 Nov 2026
Jev, TypeSafe's hosted judge model, answers bounded questions with probabilities. mu reaches it through TypeSafe, OpenRouter, the Vercel AI Gateway, OpenCode Zen, Cloudflare Workers AI or any service that speaks the same protocol, each with its own key (TYPESA…github.com ↗ · read 18 Nov 2026
https://github.com/qybaihe/mu/releasesgithub.com ↗ · read 18 Nov 2026
<img src="docs/readme/cli-light.png" width="720" alt="mu's welcome screen in a terminal: the μ mark, "mu v0.3.8, judgment-first coding agent, built on pi 1.4.0", then the judge (the free Jev 1.13, answering in 288 ms, decision points active), the mod…github.com ↗ · read 18 Nov 2026
mude: re-encodes the whole state once per question; a shared-prefix cache would cut multi-question requests ~16-40× (skills.disclosure / tool.admission hit judge timeouts)api.github.com ↗ · read 18 Nov 2026
Feature request: a built-in consumer for session distillation on session_shutdownapi.github.com ↗ · read 18 Nov 2026
- Terminal-Bench 2.1 through Harbor 0.24.0, at each task's own time limit. Harbor runs a task's tests even when the agent runs out of time, so an attempt can hit the limit and still pass. pytorch-model-recovery is left out of the totals, because the harness di…github.com ↗ · read 18 Nov 2026
- MiMo v2.6 Pro with reasoning effort high, everywhere. pi 1.4.0 is the control. mu is mu-agent 0.1.12, built on pi 1.3.0, with its defaults; Jev is version 1.14. The mude column is a pre-release build with 24 layers, served on the same Mac. mude 1.0 is the 20…github.com ↗ · read 18 Nov 2026
A hive is two to six bees, each with its own focus. Bees read code, run commands and browse the web, but never edit; the main model makes the changes. Each time a bee finishes a message, `hive.publish` asks once: is there a finding, a dead end, a decision or a…github.com ↗ · read 18 Nov 2026
The board is append-only, so an earlier conclusion stays even after a later one overturns it: a bee reports that the tests won't run, then unsets an environment variable and they do. `hive.relate` decides how two notes relate: *supersedes*, *contradicts* or *s…github.com ↗ · read 18 Nov 2026
Terminal-Bench 2.1 through Harbor 0.24.0, at each task's own time limit. Harbor runs a task's tests even when the agent runs out of time, so an attempt can hit the limit and still pass. pytorch-model-recovery is left out of the totals, because the harness didn…github.com ↗ · read 18 Nov 2026
MiMo v2.6 Pro with reasoning effort high, everywhere. pi 1.4.0 is the control. mu is mu-agent 0.1.12, built on pi 1.3.0, with its defaults; Jev is version 1.14. The mude column is a pre-release build with 24 layers, served on the same Mac. mude 1.0 is the 20-l…github.com ↗ · read 18 Nov 2026
<p align="center"><code>npm i -g mu-agent && mu setup && mu</code> · Node.js 22.19 or newer · or the <a href="https://github.com/qybaihe/mu/releases/latest">desktop app</a> for macOS, Windows and Linux</p>raw.githubusercontent.com ↗ · read 18 Nov 2026
mu (μ): most of what a coding agent decides isn't code.api.github.com ↗ · read 18 Nov 2026
609api.github.com ↗ · read 18 Nov 2026
From the API response. No sentence to quote.api.github.com ↗ · read 18 Nov 2026
From the API response. No sentence to quote.api.github.com ↗ · read 18 Nov 2026
From the API response. No sentence to quote.api.github.com ↗ · read 18 Nov 2026
From the API response. No sentence to quote.github.com ↗ · read 18 Nov 2026
Ordered by Index score.
Terminal coding agent
Terminal coding agent
Terminal coding agent
Terminal coding agent
Terminal coding agent
Terminal coding agent
Terminal coding agent
Terminal coding agent