adk-rust-realtime-voice

star 464

Implement realtime voice agents with ADK-Rust, including audio formats, event handling, and streaming reliability. Use when building or debugging voice interactions in adk-realtime.

zavora-ai By zavora-ai schedule Updated 3/29/2026

name: adk-rust-realtime-voice description: Implement realtime voice agents with ADK-Rust, including audio formats, event handling, and streaming reliability. Use when building or debugging voice interactions in adk-realtime.

ADK Rust Realtime Voice

Overview

Build low-latency bidirectional audio flows with explicit session/event handling.

Workflow

  1. Select provider backend and audio format early.
  2. Validate session configuration and VAD settings.
  3. Handle client/server events with explicit match branches.
  4. Verify tool call lifecycle in realtime paths.
  5. Configure interruption detection (Manual or Automatic).
  6. Use SessionUpdateConfig for mid-session context mutation (swap instructions/tools without dropping the call).

Guardrails

  1. Fail fast on unsupported modalities.
  2. Keep audio encoding assumptions explicit.
  3. Add coverage for cancellation, interruptions, and error events.
  4. Use ContextMutationOutcome to handle provider differences (OpenAI: Applied, Gemini: RequiresResumption).

References

  • Use references/realtime-voice-playbook.md.
Install via CLI
npx skills add https://github.com/zavora-ai/adk-rust --skill adk-rust-realtime-voice
Repository Details
star Stars 464
call_split Forks 68
navigation Branch main
article Path SKILL.md
More from Creator