ilia/atlas
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
atlas/tickets/backlog/TICKET-010_streaming-asr-service.md
ilia 7c633a02ed Initialize project structure with essential files and documentation 
- Added .cursorrules for project guidelines and context
- Created README.md for project overview and goals
- Established ARCHITECTURE.md for architectural documentation
- Set up tickets directory with initial ticket management files
- Included .gitignore to manage ignored files and directories

This commit lays the foundation for the Atlas project, ensuring a clear structure for development and collaboration.
2026-01-05 20:09:44 -05:00

1.4 KiB
Raw Blame History

Ticket: Implement Streaming Audio Capture → ASR Service

Ticket Information

  • ID: TICKET-010
  • Title: Implement Streaming Audio Capture → ASR Service
  • Type: Feature
  • Priority: High
  • Status: Backlog
  • Track: Voice I/O
  • Milestone: Milestone 2 - Voice Chat MVP
  • Created: 2024-01-XX

Description

Build streaming ASR service:

  • Implement audio capture (GStreamer/ffmpeg or browser getUserMedia)
  • Create WebSocket endpoint for audio streaming
  • Integrate selected ASR engine
  • Stream audio chunks to ASR and return text segments
  • Handle start/stop based on wake-word events

Acceptance Criteria

  • Audio capture working (from mic or WebSocket)
  • WebSocket endpoint for audio streaming
  • ASR engine integrated
  • Text segments returned with timestamps
  • Handles wake-word start/stop events
  • Streaming latency acceptable (< 2s end-to-end)

Technical Details

Implementation:

  • Audio capture: PyAudio, GStreamer, or browser MediaRecorder
  • WebSocket server for real-time streaming
  • ASR processing: faster-whisper or selected engine
  • Return format: JSON with text, timestamps, confidence

Dependencies

  • TICKET-009 (ASR engine selection)
  • Wake-word event flow defined (from TICKET-006)

Related Files

  • home-voice-agent/asr/ (to be created)

Notes

Can run in parallel with TTS and LLM work. Needs wake-word integration for start/stop.