How daxeshchaudhary/chatgpt-agent works
AI Learning Studio takes a question from a web interface or command-line entry point, identifies its topic, gathers relevant web material and video recommendations, and builds or reuses a topic-specific retrieval database. It retrieves context to draft and review study notes, then returns the notes with video and image recommendations. The graph follows the sampled pipeline and tool calls; template contents, configuration, and some implementation details were not sampled.
Main entry points: pipeline.py Learning Pipeline · rag.py RAG Service · tools.py Research Tools · app.py FastAPI App
Components 18
User Interface
- Study Studio
static/script.js - Studio Page
templates/index.html
Learning Workflow
- FastAPI App
app.py - CLI Entry
main.py - Learning Pipeline
pipeline.py
Research and Retrieval
- Search Agent
search_agent.py - Content Agent
content_agent.py - RAG Service
rag.py - Research Tools
tools.py - Topic Vector Store
Teaching and Review
- Teaching Agent
teaching_agent.py - Critic Agent
critic_agent.py - LLM Client
llm.py
External Services
- Mistral API
- Tavily Search
- YouTube Search
- Embedding Model
Other components
- Learner
Connections 28
- Learner → Study Studio: asks question
- Study Studio → FastAPI App: posts question
- FastAPI App → Studio Page: renders page
- FastAPI App → Learning Pipeline: runs workflow
- Learning Pipeline → FastAPI App: returns results
- FastAPI App → Study Studio: returns JSON
- Learner → CLI Entry: enters question
- CLI Entry → Learning Pipeline: runs workflow
- Learning Pipeline → Teaching Agent: identifies topic
- Learning Pipeline → RAG Service: checks topic store
- Learning Pipeline → Search Agent: searches question
- Search Agent → Research Tools: searches web and video
- Research Tools → Tavily Search: searches web
- Research Tools → YouTube Search: searches videos
- Learning Pipeline → Content Agent: extracts web sources
- Content Agent → Research Tools: extracts webpages
- Learning Pipeline → RAG Service: indexes documents
- RAG Service → Topic Vector Store: creates and persists
- RAG Service → Embedding Model: embeds documents
- Learning Pipeline → RAG Service: retrieves context
- RAG Service → Topic Vector Store: loads and searches
- Learning Pipeline → Teaching Agent: drafts and revises notes
- Learning Pipeline → Critic Agent: requests review
- Critic Agent → LLM Client: generates review
- Teaching Agent → LLM Client: generates notes
- LLM Client → Mistral API: sends prompts
- Learning Pipeline → Research Tools: searches topic images
- Research Tools → Tavily Search: searches images