[Suggestion] Deno Deploy serverless audio transcription (TypeScript)

[deepgram-robot]

What to build

A serverless audio transcription API deployed to Deno Deploy, using Deepgram's pre-recorded and streaming STT APIs with Deno-native patterns (no Node.js compatibility layer).

Why this matters

Deno Deploy is growing as a serverless platform that offers globally distributed edge execution with zero cold starts. Developers choosing Deno for its security model, TypeScript-first design, and web-standard APIs need a reference implementation for audio transcription. While Deepgram has Deno starters in deepgram-starters, there is no deployed serverless example showing how to build a production-ready transcription API endpoint on Deno Deploy with proper error handling, CORS, and streaming support.

Suggested scope

• Language: TypeScript (Deno-native, no npm: specifiers where avoidable)
• Platform: Deno Deploy
• Deepgram APIs: STT (pre-recorded for file upload, streaming for real-time)
• What it should include:
  • REST endpoint for pre-recorded audio transcription (file upload or URL)
  • WebSocket endpoint for real-time streaming transcription
  • Deno-native fetch and WebSocket APIs (no Node.js polyfills)
  • CORS middleware for browser clients
  • Simple HTML frontend for testing
  • deno.json configuration for Deno Deploy
• Complexity: Low-medium

Acceptance criteria

• Runnable with minimal setup (clone, add API key, deno task dev)
• README explains Deno Deploy deployment process
• Uses Deno-native APIs without Node.js compatibility layer
• Includes both pre-recorded and streaming transcription endpoints
• Deployable to Deno Deploy with one command

---

Raised by the DX intelligence system.