What to build
A working example showing how to build a Vapi.ai voice agent that uses Deepgram as the STT and TTS provider, with custom function calling and server-side webhook handling.
Why this matters
Vapi.ai is one of the most popular voice agent platforms, and many developers building production voice agents use Vapi as their orchestration layer. Developers need a clear reference showing how to configure Deepgram as the speech provider in Vapi, handle server events, and implement custom tool calls. This helps developers who want Deepgram's accuracy in a managed voice agent platform.
Suggested scope
- Language: TypeScript (Node.js)
- Framework: Vapi Server SDK + Express
- Deepgram APIs: STT (Nova-3) and TTS (Aura) via Vapi provider config
- What it does: Creates a Vapi assistant configured with Deepgram STT/TTS, implements a webhook server for function calling (e.g., lookup customer info), handles call events
- Complexity: Medium — server + Vapi config
Acceptance criteria
Raised by the DX intelligence system.
What to build
A working example showing how to build a Vapi.ai voice agent that uses Deepgram as the STT and TTS provider, with custom function calling and server-side webhook handling.
Why this matters
Vapi.ai is one of the most popular voice agent platforms, and many developers building production voice agents use Vapi as their orchestration layer. Developers need a clear reference showing how to configure Deepgram as the speech provider in Vapi, handle server events, and implement custom tool calls. This helps developers who want Deepgram's accuracy in a managed voice agent platform.
Suggested scope
Acceptance criteria
Raised by the DX intelligence system.