How vishesh162/video-lens works
Video Lens turns an uploaded recording or public YouTube link into a transcript, structured notes, transcript-grounded chat, and a downloadable PDF. The main path runs from the workspace through the HTTP job API and Python analysis pipeline; transcript retrieval and note generation are distinct stages. The graph also preserves the documented walkthrough and local YouTube setup. Some frontend-to-API details and unsampled analysis capabilities are documented rather than verified from source excerpts, so those links cite the README.
Main entry points: workspace.tsx Analysis Workspace · main.py Analysis Orchestrator · rag_engine.py Transcript Q&A · api_server.py HTTP Job API
Components 24
Workspace UI
- Analysis Workspace
frontend/src/components/workspace.tsx - API Client
frontend/src/lib/api.ts - Notes and Chat
frontend/src/components/notes-panel.tsx - Assistant Chat
frontend/src/components/assistant-chat.tsx - Video Player
frontend/src/components/video-player.tsx - PDF Export
frontend/src/lib/notes.ts - Prepared Walkthrough
frontend/src/lib/walkthrough.ts - Local YouTube Guide
frontend/src/components/youtube-local-guide.tsx
HTTP and Jobs
- HTTP Job API
api_server.py - Streamed Upload
utils/upload_stream.py
Recording Analysis
- Analysis Orchestrator
main.py - Audio Preparation
utils/audio_processor.py - Transcription
core/transcriber.py
Notes and Retrieval
- Summary and Title
core/summarizer.py - Structured Findings
core/extractor.py - Transcript Q&A
core/rag_engine.py - Lexical Retrieval
core/transcript_search.py - Vector Retrieval
core/vector_stores.py - Chat Model Adapter
core/llm.py
Other components
- Recording User
- Whisper Service
- Sarvam Service
- Configured LLM Provider
- YouTube
Connections 26
- Recording User → Analysis Workspace: uses
README.md - Analysis Workspace → API Client: requests analysis
README.md - API Client → HTTP Job API: calls API
README.md - Analysis Workspace → Notes and Chat: shows notes
frontend/src/components/workspace.tsx - Analysis Workspace → Assistant Chat: shows chat
frontend/src/components/workspace.tsx - Analysis Workspace → Video Player: shows playback
frontend/src/components/workspace.tsx - Notes and Chat → PDF Export: exports PDF
- Analysis Workspace → Prepared Walkthrough: offers walkthrough
README.md - Recording User → Local YouTube Guide: opens setup guide
README.md - HTTP Job API → Streamed Upload: streams recording
api_server.py - HTTP Job API → Analysis Orchestrator: runs analysis
api_server.py - Analysis Orchestrator → Audio Preparation: prepares source
main.py - Analysis Orchestrator → Transcription: transcribes chunks
main.py - Analysis Orchestrator → Summary and Title: generates notes
main.py - Analysis Orchestrator → Structured Findings: extracts findings
main.py - Analysis Orchestrator → Transcript Q&A: builds Q&A
main.py - Audio Preparation → YouTube: downloads audio
utils/audio_processor.py - Transcription → Whisper Service: transcribes English
core/transcriber.py - Transcription → Sarvam Service: transcribes Hinglish
core/transcriber.py - Summary and Title → Chat Model Adapter: uses model
core/summarizer.py - Structured Findings → Chat Model Adapter: uses model
core/extractor.py - Transcript Q&A → Chat Model Adapter: generates answers
core/rag_engine.py - Transcript Q&A → Lexical Retrieval: retrieves passages
core/rag_engine.py - Transcript Q&A → Vector Retrieval: uses optional store
core/rag_engine.py - Chat Model Adapter → Configured LLM Provider: calls configured API
- HTTP Job API → Transcript Q&A: answers chat
api_server.py