Tech Stack and Summary

Technology stack

Backend

• Framework: Python 3.10+ / FastAPI + Pydantic v2
• Cognitive engine: the in-house CogArch 2.0 architecture
• LLM access: LLM Provider management (35+ providers + OpenCode)
• Persistence: SQLite (FTS5 full-text search + WAL) + LSM-Tree storage for the memory layer
• Vector retrieval: FAISS + BGE embedding models (local ONNX inference)
• Event system: an in-house EventBus message bus
• Authentication: JWT (PyJWT) + bcrypt password hashing (passlib)
• Testing: pytest (846 test files — 703 unit, 54 integration)

Frontend

• Framework: Vue 3 + TypeScript + Vite + Pinia + Ant Design Vue
• Testing: Vitest (42 files, 614 cases all green)
• Internationalization: 11 languages aligned key by key

Rust crates (partial)

• neurova-agent / events / isolation / loop / memory / skills / tools (version 1.0.0)

Summary

Neurova is a complete cognitive multi-agent AI framework. It:

• is not just a chatbot — it is an autonomous agent with its own memory, personality and capacity to grow
• is not a single-agent system — it is a platform where multiple agents collaborate in parallel
• is not a black-box LLM wrapper — it is an intelligent engine with a complete closed-loop cognitive architecture
• is not a tool that ignores security — it is a production-grade security system with a built-in layered firewall and MCP governance

Core strengths

Quick start

# 2. Install dependencies
pip install -r requirements.txt
# 3. Configure the environment
cp .env.example .env
# Edit .env and fill in your API key
# 4. Start the backend (port 9527)
python start.py --backend
# 5. Start the frontend (port 8100)
cd NeurUI && npm install && npm run dev
# 6. Start everything
python start.py --chat
python start.py --check
```
## Project links
- **GitHub**: https://github.com/kingsa2026/Neurova
- **Documentation**: https://www.neurova.top/docs
---
*Neurova (Zhixing) — giving every star some warmth.*