Search Agent Video Demo
🚀 Deep Search AI — Local-First Research → Action System
Deep Search AI is a powerful, privacy-focused AI research and automation system that connects search, reasoning, and real-world actions into a single seamless workflow.
This demo showcases:
• 🔎 Multi-provider Web Search (SearxNG / Tavily / SerpAPI)
• 🧠 RAG + CAG (Retrieval-Augmented + Cache-Augmented Generation)
• ⚡ Smart Caching for instant repeated queries
• 🎯 Perspective Dial for controlled information bias
• 🧑💻 Assistant Personas (Research → Action bridge)
• 📂 Local Executor for safe automation
• 🔐 Config-driven, self-hostable architecture
• 🐳 Docker-based deployment
⸻
🧠 What Makes This Different?
Most AI tools stop at answers.
Deep Search AI goes further:
• Turn research into actions instantly
• Use a self-hosted search backend (SearxNG)
• Keep control of your data
• See why decisions were made (Explain Mode)
• Run it locally or deploy via Docker
Search → Understand → Act.
⸻
🛠 Tech Stack
• FastAPI (Backend)
• Next.js (Frontend)
• Rust Executor (Secure local actions)
• Redis (Caching)
• SearxNG (Self-hosted web search)
• Docker Compose deployment
⸻
📦 GitHub Repository
👉 https://github.com/somsubhra88/deep_search_ai
Feel free to star ⭐ the repo if you find it useful!
⸻
🔮 Roadmap
• Advanced persona system
• Command palette (⌘K)
• Semantic caching improvements
• Self-host templates (Coolify / Dokploy)
• Enterprise audit layer
⸻
If you’re interested in:
• Building your own AI research engine
• Running a private AI system at home
• Creating autonomous but safe assistants
This project is for you.