XiaomiMiMo/MiMo-Code

MiMo Code: Where Models and Agents Co-Evolve

What it solves

MiMoCode is a terminal-native AI coding assistant designed to handle complex, long-horizon software development tasks. It solves the problem of losing project context across sessions and the inefficiency of manual, repetitive coding workflows by providing persistent memory and structured agentic orchestration.

How it works

It operates as a CLI tool that connects to various LLM providers. The system uses a multi-agent architecture with specialized agents (Build, Plan, Compose) and a subagent system for parallel execution. To maintain context, it employs a persistent memory system using SQLite FTS5, creating project-level memory, session checkpoints, and task progress logs. It also features a "Dream & Distill" mechanism to extract knowledge and automate repeated patterns into reusable skills.

Who it’s for

Developers who prefer terminal-based workflows and want an AI assistant that can manage Git, run commands, and maintain a deep, long-term understanding of their codebase across multiple sessions.

Highlights

  • Persistent Memory: Uses MEMORY.md and automatic checkpoints to ensure agents don't have to relearn project context.
  • Multi-Agent Orchestration: Specialized modes for building, planning, and specs-driven development (Compose mode).
  • Deterministic Workflows: JavaScript-based scripts for complex pipelines like deep research, fact-checking, and autonomous optimization loops.
  • Extensible Skill System: A library of built-in skills (e.g., arXiv search, PDF/PPTX manipulation) that can be overridden or custom-created.
  • Intelligent Context Management: Budgeted token injection and adjustable compaction points to optimize cost and performance.
  • Voice Input: Real-time streaming voice transcription for hands-free interaction.

Related

  • Project
    LiuMengxuan04/MiniCode

    A lightweight terminal coding assistant that implements a Claude Code-like agentic workflow for local development, focusing on simplicity and extensibility.

  • Project
    memcode-in/memcode

    An end-to-end memory infrastructure for AI that provides agents and applications with persistent, selective, and queryable memory across sessions and model changes.

  • Project
    tickernelz/opencode-mem

    A persistent memory system for AI coding agents that uses a local vector database to retain long-term project context and user preferences across sessions.

  • Project
    shareAI-lab/Kode-CLI

    A terminal-based AI coding assistant that integrates codebase understanding, file editing, and command execution with support for multi-model collaboration and sub-agents.

  • Project
    akitaonrails/ai-memory

    ai‑memory is a Rust‑based server that records every observation a coding‑assistant makes (prompts, tool calls, session boundaries) as plain markdown files in a git‑backed wiki. The server can be run locally, in Docker/Podman, or on a homelab box and is designed to be shared across many agents (Claude Code, Codex, Cursor, Gemini CLI, etc.), machines and team members. Hooks posted by agents are consolidated into readable wiki pages at session end; the next session receives a concise handoff so it can continue exactly where the previous one stopped. Storage is human‑readable markdown plus a SQLite index (FTS5, entity and optional vector search). By default it needs no LLM or embedding keys; optional Anthropic/OpenAI integrations add LLM‑generated summaries and semantic search. Multi‑user authentication, audit logs and per‑person attribution are built in. Install via Arch AUR, Docker wrapper, or native binary, start the server, register a client and its hooks, and the system works transparently.