Skip to content

Support advanced prompt voice agent settings - #9655

Draft
JianW (v1212) wants to merge 9 commits into
Azure:mainfrom
v1212:users/wujia/prompt-voice-agent-advanced-settings-unified
Draft

Support advanced prompt voice agent settings#9655
JianW (v1212) wants to merge 9 commits into
Azure:mainfrom
v1212:users/wujia/prompt-voice-agent-advanced-settings-unified

Conversation

@v1212

@v1212 JianW (v1212) commented Aug 20, 2026

Copy link
Copy Markdown
Contributor

Summary

  • Supersedes Support advanced prompt voice agent configuration #9609 by rebasing advanced prompt voice customization on top of Add opt-in unified API mode for prompt voice agents #9634's unified /agents API mode.
  • Adds editable azure.yaml advanced prompt voice settings for structured inputs, richer audio input/output, output modalities, tools, avatar, greeting, handoff, and response/tool options.
  • Keeps legacy/default prompt voice behavior protected: advanced settings require AZURE_VOICE_AGENT_API=unified-flat so unsupported legacy deployments fail clearly instead of silently dropping fields.

Dependency

Vienna / sample UI alignment

  • Aligns with Vienna prompt voice validation and E2E coverage for structured_inputs, audio format/VAD/transcription, noise reduction, tools (function, mcp, system, toolbox), avatar, store, and output modalities.
  • Aligns with the sample UI create contract for avatar, greeting, tools, store, and audio defaults.
  • Uses the unified flat output voice wire shape in unified-flat mode while keeping YAML authoring friendly via audio.output.voice.type/name/style/....

Validation

  • go test ./... -count=1
  • go build ./...
  • cspell lint 'extensions/**/*.go' 'extensions/**/*.md' --config ./.vscode/cspell.yaml --no-progress
  • Live West US 2 TiP managed advanced smoke with AZURE_VOICE_AGENT_API=unified-flat: provision, deploy, GET /agents/{name} assertions for structured inputs, azure_semantic_vad, transcription, flat output voice shape, output modalities, system tool, avatar, and greeting; resources cleaned up.
  • Live West US 2 TiP guide-style advanced validation with managed, BYOM realtime, and BYOM cascaded agents. All three deployed with advanced YAML and passed GET /agents/{name} assertions for model type/model id, structured inputs, azure_semantic_vad, transcription, flat output voice shape, output modalities, system tool, avatar, and greeting; resources cleaned up.

@azure-pipelines

Copy link
Copy Markdown
Azure Pipelines:
Successfully started running 1 pipeline(s).
20 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

@v1212
JianW (v1212) force-pushed the users/wujia/prompt-voice-agent-advanced-settings-unified branch from ab7cd25 to 5d06206 Compare August 21, 2026 05:39
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ext-agents azure.ai.agents extension

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants