Build and deploy AI agents from your own data in under 60 seconds

Hacker News

Share card

Build and deploy AI agents from your own data in under 60 seconds

Hi HN, I built Bot The Builder to solve a recurring problem I kept facing while building chatbots for clients: every time, I had to rebuild the same RAG stack, embed the same knowledge base, and deploy from scratch—even for small use cases. So I created a tool that lets you build and deploy AI agents from your own data (PDFs, docs, URLs, text) in under 60 seconds — no code required. You can embed the agent in any website or access it via API. ### Why I built it: - Most “AI chatbot builders” are either locked down or require complex LangChain pipelines. - I needed something fast, modular, and production-ready that doesn’t hide the underlying architecture. - This is a developer-first tool, not a marketing product. ### How it works: 1. Upload your own content or paste a URL 2. The system creates embeddings (using OpenAI or local models, FAISS vector store) 3. You instantly get a deployed agent with memory + context 4. You can access the agent via the url: https://botthebuilder.net/hosted-chatbot/love.com if built for love.com or embed it directly on your app/site ### Tech stack: - Vector DB: FAISS - Embeddings: OpenAI + Local model support - Frontend: React + Vite - Backend: Python / FastAPI - Architecture is modular — not tied to a single provider ### Use cases I'm seeing: - Internal knowledge assistants for startups - Agencies building AI bots for client websites - Solo founders launching niche AI products - Automated onboarding, support, or documentation agents ### Live demo: https://BotTheBuilder.net It's still early, so I'm looking for feedback from builders here: - What features would make it actually useful in your workflow? - Would you want more control over the agent’s chain-of-thought/system prompt? Happy to answer any technical questions about the architecture or roadmap. Feedback and criticism are very welcome.