[Suggestion] Inngest serverless audio transcription pipeline with webhook callbacks (TypeScript)

[deepgram-robot]

What to build

An Inngest-powered serverless pipeline that processes audio files through Deepgram's STT and Audio Intelligence APIs, triggered by events (file upload, webhook, or schedule). The pipeline handles long-running transcription jobs with automatic retries, step-level caching, and webhook callbacks on completion.

Why this matters

Inngest is a rapidly growing serverless background job platform that solves the "too long for a Lambda" problem. Audio transcription of long files often exceeds serverless function timeout limits. Inngest's step functions with built-in retries and caching make it ideal for audio processing pipelines. Developers building event-driven audio workflows on Vercel, Netlify, or standalone Node.js need a reference implementation that combines Inngest's durability with Deepgram's transcription APIs.

Suggested scope

• Language: TypeScript
• Deepgram APIs: STT pre-recorded, Audio Intelligence
• Framework: Inngest SDK
• Deployment: Vercel, Netlify, or standalone Node.js
• Key patterns: Multi-step function, event triggering, webhook callbacks, step caching
• Complexity: Moderate

Acceptance criteria

• Runnable with minimal setup (clone, add API key, run)
• README explains the pattern clearly
• Uses current SDK version
• Multi-step Inngest function with individual step retries
• Supports event-triggered and scheduled execution
• Sends webhook callback on completion with transcription results
• Includes deployment instructions for at least one platform

---

Raised by the DX intelligence system.