ResearchBot

A general-purpose AI research platform that takes a location + topic, browses the web for real-world data, generates summaries, and stores everything as a knowledge graph.

Work in progress.

Goal

Build a knowledge acquirer agent that can research specific queries and store structured findings into a graph database. Example: "What are good venues for musicians to reach out to in Middlesbrough?" → list of venues with contact details, live music ratings, and recent gig data.

Stack

• Agent — TypeScript browser app (Vite)
• Server — Go REST API, tool execution, graph writes
• Database — Neo4j knowledge graph
• LLM — llama-server (local inference)
• Diagrams — PlantUML server (Docker, port 7070)

Diagrams

PlantUML diagrams live in docs/. Render with:

curl -s http://localhost:7070/png -H 'Content-Type: text/plain' \
  --data-binary @docs/your-diagram.puml \
  --output docs/your-diagram.png

Status

Foundation phase complete. Knowledge graph schema, research pipeline, and tool integration in progress.