An open-source self-evolving AI agent framework with multi-persona, multi-LLM endpoint, and multi-IM channel support, featuring daily self-check and automatic skill generation.
OpenAkita is an open-source self-evolving AI agent framework.
Core Features#
Self-Evolution Capabilities#
- Daily Self-Check (03:00): Memory consolidation, semantic deduplication, insight extraction, refresh MEMORY.md
- Self-Diagnosis (04:00): Analyze error logs → LLM diagnosis → auto-fix → generate report
- Task Review: Analyze efficiency after each task, extract experience, store in long-term memory
- Auto Skill Generation: Automatically generate Skills + install dependencies when encountering bottlenecks
Multi-Persona & Interaction#
- 8 preset personas: Default Assistant, Business Consultant, Tech Expert, Personal Butler, Virtual Girlfriend, Virtual Boyfriend, Family, Jarvis
- Proactive greetings: Remember user birthday, auto-mute at night, learn user preferences
- Sticker system: 5700+ stickers, emotion sensing, persona matching
Multi-LLM Endpoint Support#
Supports 9+ providers: Tongyi Qianwen, Zhipu AI, Baidu Qianfan, DeepSeek, Kimi, OpenAI, Anthropic, Google Gemini, Groq, Mistral, with hot-swap, priority scheduling, auto-failover, and health checks.
IM Channel Support#
| Channel | Public IP Required |
|---|---|
| Telegram | No |
| Feishu | No |
| WeCom | Yes |
| DingTalk | No |
| QQ Official Bot | No |
| OneBot | No |
Desktop Terminal#
Cross-platform desktop app (Tauri + React) with streaming output, Markdown rendering, multimodal input, Thinking display, Plan mode, bilingual support (ZH/EN), status monitoring, and system tray.
Architecture#
Desktop App (Tauri + React)
Identity ─── SOUL.md · AGENT.md · USER.md · MEMORY.md · 8 Persona Presets
Core ─── Brain(LLM) · Memory(Vector) · Ralph(Never-Give-Up Loop)
Tools ─── Shell · File · Web · Browser · Desktop · MCP · Skills
Evolution ── SelfCheck · Generator · Installer · LogAnalyzer
Channels ─── CLI · Telegram · Feishu · WeCom · DingTalk · QQ Official · OneBot
Installation#
Desktop App (Recommended): Download from GitHub Releases → Quick Setup → Add LLM endpoint → Start Setup
pip Install:
pip install openakita[all]
openakita init
openakita
Source Install:
git clone https://github.com/openakita/openakita.git
cd openakita
python -m venv venv && source venv/bin/activate
pip install -e ".[all]"
openakita init
Core Commands#
openakita # Interactive chat
openakita run "Build a calculator" # Execute single task
openakita serve # Service mode (IM channels)
openakita daemon start # Background daemon
openakita status # Check status
System Requirements#
- OS: Windows 10/11 (x86_64), macOS 12+, Linux (x86_64)
- Disk: At least 2 GB
- At least one available LLM API endpoint
Tech Stack#
Python (80.6%), TypeScript (12.6%), Rust (2.9%) | MIT License | v1.24.0