A full-stack RAG project for personal introduction and project experience Q&A. The system retrieves relevant information from resume, projects, and FAQ documents to generate natural language answers.
Live Demo: [https://self-agent-web.onrender.com]
Free tier: first visit may take 30-50 seconds to wake up the backend service.
- Frontend: React + Vite + TypeScript
- Backend: FastAPI + Pydantic
- Database: PostgreSQL + pgvector
- Retrieval: Hybrid (ILIKE + Vector) + RRF fusion ranking
- LLM / Embedding: Gemini API (gemini-2.5-flash / gemini-embedding-001)
- Deployment: Render (Static Site + Web Service + PostgreSQL)
docs/— Architecture and documentationdata/— Knowledge base (profile / projects / skills / faq)backend/— API and RAG core logicfrontend/— Chat interfaceinfra/— Database init scripts (init.sql / docker-compose)
- Architecture:
docs/architecture.md
For learning and personal project demonstration only.
一个面向"个人介绍与项目经历问答"的 RAG 全栈项目。用户提问后,系统基于简历、项目与 FAQ 资料进行检索增强生成,返回自然语言回答。
在线体验:[https://self-agent-web.onrender.com]
Free tier 首次访问需等待 30-50 秒唤醒后端服务。
- 前端:React + Vite + TypeScript
- 后端:FastAPI + Pydantic
- 数据库:PostgreSQL + pgvector
- 检索策略:Hybrid(ILIKE + Vector)+ RRF 融合排序
- LLM / Embedding:Gemini API(gemini-2.5-flash / gemini-embedding-001)
- 部署:Render(前端 Static Site + 后端 Web Service + PostgreSQL)
docs/:架构与说明文档data/:知识库数据(profile / projects / skills / faq)backend/:API 与 RAG 核心逻辑frontend/:聊天界面infra/:数据库初始化脚本(init.sql / docker-compose)
- 架构文档:
docs/architecture.md
仅用于学习与个人项目演示。