LibreAI / Dev Tools / MiMo-Code
// devtools · XiaomiMiMo/MiMo-Code

MiMo-Code

Xiaomi's terminal-native AI coding assistant where models and agents co-evolve, with project memory.

★ 13.6k stars⑂ 1.4k forksMITupdated 2026-10
▦ visual — this project's README ships no embeddable screenshot,
so here's a placeholder. The guide below is complete regardless.
// project facts

Project facts

MiMo-Code is an open-source dev tools project by XiaomiMiMo (XiaomiMiMo/MiMo-Code on GitHub). It has 13,614 GitHub stars and 1,424 forks, is released under the MIT license, and was last updated on GitHub on 8 October 2026.

GitHub stars
13,614
Forks
1,424
License
MIT
Last updated on GitHub
8 October 2026
Category
Dev Tools
Listing
Full install guide
Official site
Official site
// how it works

How it works

  1. Install the CLI

    Use the one-line installer for your platform or npm install -g @mimo-ai/cli, then run mimo.

  2. Connect a model provider

    On first launch, log in with Xiaomi MiMo or Codex OAuth, import your Claude Code authentication, or add a provider API key or custom OpenAI-compatible endpoint.

  3. Start in your project

    Open mimo in a project directory and describe a task; switch between build, plan and compose agents with Tab as the work demands.

  4. Let memory compound

    As you work, MiMo-Code writes checkpoints and project memory automatically, so the next session resumes with context instead of starting cold.

// features

Why people love it

Multiple primary agents

build (full tool permissions), plan (read-only analysis and design) and compose (specs-driven orchestration) agents, switchable with Tab, plus on-demand subagents that work in parallel.

Persistent cross-session memory

Project memory (MEMORY.md), automatic session checkpoints, scratch notes and per-task progress, powered by SQLite FTS5 full-text search and injected when a session resumes.

Intelligent context management

Automatic checkpoints, context reconstruction from memory when the window fills, budgeted injection with importance ranking, and an adjustable compaction point per model.

Goal and stop conditions

The /goal command sets a stopping condition that an independent judge model evaluates before the agent may stop, preventing premature 'optimistic stops' in autonomous work.

Deterministic workflows

JavaScript workflows orchestrate multiple agents in fixed phase sequences with bounded retries - built-ins include compose (spec to merged code), deep-research, fact-check and research-experiment.

Built-in skills

Reusable skill instruction sets (arXiv search, document generation, data analytics and more) are matched and loaded automatically, and you can add your own.

Advertisement
// run it

Run it yourself

  1. Clone the repository

    Clone MiMo-Code if you want the source; most users install the released CLI instead.

    git clone https://github.com/XiaomiMiMo/MiMo-Code.git
    cd MiMo-Code
  2. Install the CLI

    Install MiMo-Code with the one-line installer (macOS/Linux; a PowerShell installer exists for Windows).

    curl -fsSL https://mimo.xiaomi.com/install | bash
  3. Or install via npm, then run

    Alternatively install the package globally from npm - available on all platforms - and launch it in your project.

    npm install -g @mimo-ai/cli
    mimo
// use cases

What you can build

Advertisement
// faq

Questions, answered

Who maintains MiMo-Code?

It is developed by Xiaomi's MiMo team and open-source contributors, under the slogan 'where models and agents co-evolve'.

Is MiMo-Code free to use?

The CLI is MIT-licensed and free. Model usage goes through the provider you connect - Xiaomi MiMo, a ChatGPT/Codex subscription via OAuth, or your own API keys.

Does it need a paid API key?

You need some model access: OAuth login to MiMo or Codex, imported Claude Code auth, a provider API key, or any OpenAI-compatible endpoint you control.

How does MiMo-Code compare with Claude Code or OpenCode?

Its headline differences are the persistent Markdown memory system with checkpoints, multiple built-in primary agents, judge-verified goal conditions, and deterministic multi-agent workflows.

How do I get started?

Run the one-line installer or npm install -g @mimo-ai/cli, type mimo in a project, and complete the guided first-launch provider setup.

// keep exploring

Similar tools

Dev Tools

agav

Terminal-native AI coding assistant for real repositories.

★ 147 starsApache-2.0
View guide →
Dev Tools

memorax-code

Shared memory for coding agents: lessons, repo knowledge, and your preferences, carried across sessions.

★ 2.1k starsMIT
View guide →
Dev Tools

qwen-code

Alibaba's open-source terminal coding agent for Qwen models — a Gemini CLI-style workflow, free to run

★ 28.4k starsApache-2.0
View guide →
Dev Tools

twinny

Open-source AI coding assistant for VS Code. Code completion, chat, edits and reviews with local or hosted models. Your models, your infrastructure.

★ 3.7k starsMIT
View guide →
Dev Tools

gentle-ai

Gentle-AI configures the AI coding agents you already use: Claude Code, Cursor, OpenCode, Codex, Pi, and more. Choose persistent memory, Organic-Driven Devel…

★ 7.5k starsMIT
View guide →
Dev Tools

cavemem

Persistent memory for coding agents — frozen, but its core lives on in Caveman.

★ 678 starsMIT
View guide →
// alternatives

Alternatives in Dev Tools

The most-starred other Dev Tools projects in the LibreAI catalogue — the realistic shortlist if MiMo-Code is not the right fit.

Get MiMo-Code

Free and open — grab the source, star the repo, and start creating.