[Suggestion] Litestar real-time streaming transcription API with WebSocket (Python)

[deepgram-robot]

What to build

A working example of a Litestar (Python) web application that provides real-time streaming transcription via WebSocket — accepting audio from a browser microphone and returning live transcription results using Deepgram's streaming STT API.

Why this matters

Litestar is a high-performance, modern Python web framework that has grown rapidly as an alternative to FastAPI and Flask. It has native WebSocket support, dependency injection, and async-first design that make it well-suited for real-time audio streaming applications. Developers choosing Litestar for new projects need a reference implementation for integrating Deepgram streaming STT. Deepgram has examples for FastAPI, Flask, and Django but no Litestar example, leaving a gap in Python framework coverage.

Suggested scope

• Language: Python
• Framework: Litestar (latest stable)
• Deepgram APIs: Streaming STT (Nova-3)
• Includes: Litestar WebSocket handler, browser-side JavaScript for microphone capture, HTML page with live transcript display, proper connection lifecycle management (open/close/error)
• Complexity: Beginner — single-file application with minimal dependencies

Acceptance criteria

• Runnable with minimal setup (clone, add API key, pip install, run)
• README explains the Litestar WebSocket pattern clearly
• Uses current SDK version
• Browser microphone audio streams to server, live transcript displays in the page
• Handles WebSocket disconnection and reconnection gracefully
• Works with Litestar's native WebSocket support (no additional WebSocket libraries)

---

Raised by the DX intelligence system.