Self-Evolving AI Agent — Learns Autonomously, Never Gives Up

Desktop Terminal • Features • Quick Start • Architecture • Documentation

English | 中文

OpenAkita is a self-evolving AI Agent framework. It autonomously learns new skills, performs daily self-checks and repairs, accumulates experience from task execution, and never gives up when facing difficulties — persisting until the task is done.

Like the Akita dog it's named after: loyal, reliable, never quits.

• Self-Evolving — Auto-generates skills, installs dependencies, learns from mistakes
• Never Gives Up — Ralph Wiggum Mode: persistent execution loop until task completion
• Growing Memory — Remembers your preferences and habits, auto-consolidates daily
• Standards-Based — MCP and Agent Skills standard compliance for broad ecosystem compatibility
• Multi-Platform — Desktop Terminal GUI, CLI, Telegram, Feishu, DingTalk, WeCom, QQ

OpenAkita provides a cross-platform Desktop Terminal (built with Tauri + React) — an all-in-one AI assistant with chat, configuration, monitoring, and skill management:

• AI Chat Assistant — Streaming output, Markdown rendering, multimodal input, Thinking display, Plan mode
• Bilingual (CN/EN) — Auto-detects system language, one-click switch, fully internationalized
• Guided Setup Flow — 9-step wizard, streamlined and focused, dialog-based LLM endpoint management
• Localization & i18n — First-class support for Chinese and international ecosystems, PyPI mirrors, IM channels
• LLM Endpoint Manager — Multi-provider, multi-endpoint, auto-failover, online model list fetching
• IM Channel Setup — Telegram, Feishu, WeCom, DingTalk, QQ — all in one place
• Persona & Living Presence — 8 role presets, proactive greetings, memory recall, learns your preferences
• Skill Marketplace — Browse, download, configure skills in one place
• Status Monitor — Compact dashboard: service/LLM/IM health at a glance
• System Tray — Background residency + auto-start on boot, one-click start/stop

Download: GitHub Releases

Available for Windows (.exe) / macOS (.dmg) / Linux (.deb / .AppImage)

One of OpenAkita's most distinctive features — not just a tool, but a lifelike assistant with personality, memory, and warmth:

The Agent proactively greets you during idle time, remembers your birthday, preferences, and work habits — like a real friend.

OpenAkita offers first-class support for both Chinese and international ecosystems:

• Chinese LLM Providers — Alibaba DashScope (Qwen), Moonshot Kimi, MiniMax, DeepSeek, SiliconFlow
• Global LLM Support — Anthropic, OpenAI, Google Gemini, and more
• Chinese IM Channels — Feishu (Lark), WeCom, DingTalk, QQ native support
• PyPI Mirrors — Built-in Tsinghua TUNA and Alibaba mirrors for faster installs in China
• Full i18n — react-i18next based, auto system language detection, one-click switch

The core differentiator: OpenAkita doesn't just execute — it learns and grows autonomously.

The easiest way — graphical guided setup, no command-line experience needed:

1. Download the installer from GitHub Releases
2. Install and launch OpenAkita Desktop
3. Follow the wizard: Workspace → Python → Install → LLM Endpoints → IM Channels → Finish & Start

# Install
pip install openakita

# Install with all optional features
pip install openakita[all]

# Run setup wizard
openakita init

Optional extras: feishu, whisper, browser, windows

git clone https://github.com/openakita/openakita.git
cd openakita
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -e ".[all]"
openakita init

# Interactive CLI
openakita

# Execute a single task
openakita run "Create a Python calculator with tests"

# Service mode (IM channels)
openakita serve

# Background daemon
openakita daemon start

# Check status
openakita status

For complex tasks, enable Thinking mode by using a *-thinking model variant (e.g., claude-opus-4-5-20251101-thinking).

# .env (minimum configuration)

# LLM API (required — configure at least one)
ANTHROPIC_API_KEY=your-api-key

# Telegram (optional)
TELEGRAM_ENABLED=true
TELEGRAM_BOT_TOKEN=your-bot-token

┌─────────────────────────────────────────────────────────────────┐
│                          OpenAkita                               │
├─────────────────────────────────────────────────────────────────┤
│                                                                  │
│  ┌──────────────────── Desktop App ────────────────────────┐   │
│  │  Tauri + React · AI Chat · Config · Monitor · Skills    │   │
│  └──────────────────────────────────────────────────────────┘   │
│                              │                                   │
│  ┌──────────────────── Identity Layer ──────────────────────┐   │
│  │  SOUL.md · AGENT.md · USER.md · MEMORY.md                │   │
│  │  Personas (8 presets + user_custom)                       │   │
│  └──────────────────────────────────────────────────────────┘   │
│                              │                                   │
│  ┌──────────────────── Core Layer ──────────────────────────┐   │
│  │  Brain (LLM) · Identity · Memory · Ralph Loop             │   │
│  │  Prompt Compiler · Task Monitor                           │   │
│  │  PersonaManager · TraitMiner · ProactiveEngine            │   │
│  └──────────────────────────────────────────────────────────┘   │
│                              │                                   │
│  ┌──────────────────── Tool Layer ──────────────────────────┐   │
│  │  Shell · File · Web · MCP · Skills · Scheduler            │   │
│  │  Browser · Desktop · Plan · Profile · IM Channel          │   │
│  │  Persona · Sticker                                        │   │
│  └──────────────────────────────────────────────────────────┘   │
│                              │                                   │
│  ┌──────────────────── Evolution Engine ────────────────────┐   │
│  │  SelfCheck · Generator · Installer · LogAnalyzer          │   │
│  │  DailyConsolidator · TaskRetrospection                    │   │
│  └──────────────────────────────────────────────────────────┘   │
│                              │                                   │
│  ┌──────────────────── Channel Layer ───────────────────────┐   │
│  │  CLI · Telegram · Feishu · WeCom · DingTalk · QQ          │   │
│  └──────────────────────────────────────────────────────────┘   │
│                                                                  │
└─────────────────────────────────────────────────────────────────┘

Join our community for help, discussions, and updates:

WeChat Group Scan to join (Chinese)

WeChat — Chinese community chat Discord — Join Discord X (Twitter) — @openakita Email — zacon365@gmail.com

• Documentation — Complete guides
• Issues — Bug reports & feature requests
• Discussions — Q&A and ideas
• Star us — Show your support

• Anthropic Claude — LLM Engine
• Tauri — Cross-platform desktop framework
• browser-use — AI browser automation
• AGENTS.md Standard — Agent behavior specification
• Agent Skills — Skill standardization specification
• ZeroMQ — Multi-agent inter-process communication
• ChineseBQB — Chinese sticker/emoji pack collection

MIT License — See LICENSE

This project includes third-party skills licensed under Apache 2.0 and other open-source licenses. See THIRD_PARTY_NOTICES.md for details.

OpenAkita — Self-Evolving AI Agent, Learns Autonomously, Never Gives Up