How gauravsrc/gnani-task works
Audio Notes accepts recordings, stores them for playback, and processes them asynchronously into transcripts and English summaries. The principal flow runs from browser upload through a queued, resumable worker job to the recording page, which polls for progress and results. The README documents frontend capabilities and integrations; backend calls are supported by the indexed source references and excerpts. Frontend component-level wiring and some backend internals were not sampled, so those relationships are shown only where documented.
Main entry points: main.py HTTP API · pipeline.py Resumable pipeline · storage.py Audio storage access · models.py Recording and chunk state
Components 21
Recording experience
- Upload form
frontend/src/components/UploadForm.tsx - Recording list
frontend/src/components/RecordingList.tsx - Recording view
frontend/src/components/RecordingView.tsx - Server status
frontend/src/components/ServerStatus.tsx - App pages
frontend/src/app/page.tsx - Recording page
frontend/src/app/recordings/[id]/page.tsx
API and job control
- HTTP API
backend/app/main.py - Job worker
backend/app/worker.py
Audio processing
- Resumable pipeline
backend/app/pipeline.py - Audio conversion and chunking
backend/app/audio.py - Speech client
backend/app/gnani.py - Transcript summarizer
backend/app/summarizer.py
State and storage
- Recording and chunk state
backend/app/models.py - Database sessions
backend/app/db.py - Audio storage access
backend/app/storage.py - Application settings
backend/app/config.py
External services
- Supabase Postgres
- Supabase Storage
- Gnani speech-to-text
- Google Gemini
Other components
- Recording user
Connections 29
- Recording user → Upload form: selects audio
README.md - Upload form → HTTP API: uploads audio
README.md - HTTP API → Audio storage access: stores audio
backend/app/main.py - HTTP API → Recording and chunk state: queues recording
README.md - Job worker → Recording and chunk state: claims jobs
backend/app/worker.py - Job worker → Resumable pipeline: processes job
backend/app/worker.py - Resumable pipeline → Audio storage access: loads audio
backend/app/pipeline.py - Resumable pipeline → Audio conversion and chunking: prepares chunks
backend/app/pipeline.py - Resumable pipeline → Speech client: transcribes chunks
backend/app/pipeline.py - Speech client → Gnani speech-to-text: sends audio
README.md - Resumable pipeline → Transcript summarizer: summarizes transcript
backend/app/pipeline.py - Transcript summarizer → Google Gemini: generates summary
README.md - Resumable pipeline → Recording and chunk state: saves progress
backend/app/pipeline.py - HTTP API → Recording and chunk state: reads recording state
backend/app/main.py - HTTP API → Audio storage access: signs playback URL
backend/app/main.py - Audio storage access → Supabase Storage: accesses bucket
README.md - Database sessions → Supabase Postgres: connects
README.md - Recording and chunk state → Database sessions: uses base
backend/app/models.py - Database sessions → Application settings: uses connection settings
backend/app/db.py - Audio storage access → Application settings: uses credentials
backend/app/storage.py - Speech client → Application settings: uses service settings
backend/app/gnani.py - Transcript summarizer → Application settings: uses model settings
backend/app/summarizer.py - Recording user → App pages: opens recordings
README.md - App pages → Recording list: shows past uploads
README.md - Recording user → Recording page: opens recording
README.md - Recording page → Recording view: shows results
README.md - Recording view → HTTP API: polls progress
README.md - Recording user → Server status: sees wake notice
README.md - Server status → HTTP API: checks health
README.md