Skip to content

feat(streaming): add assemblyai-stt-3-7-realtime speech models - #210

Merged
ccampbell-aai merged 3 commits into
mainfrom
ccampbell/streaming-stt-3-7
Nov 19, 2026
Merged

ccampbell-aai merged 3 commits into
mainfrom
ccampbell/streaming-stt-3-7

Conversation

@ccampbell-aai

@ccampbell-aai ccampbell-aai commented Nov 12, 2026 •

Copy link
Copy Markdown
Contributor

Adds assemblyai-stt-3-7-realtime (note taking) and assemblyai-stt-3-7-realtime-agent (voice agents) to StreamingSpeechModel for the Universal 3.7 launch (Tue 11/17), and "manual" to StreamingMode (the client ends each turn with forceEndpoint(); streaming API side is DeepLearning #23928). Bumps to 4.43.2.

Test plan

  • jest tests/unit/streaming.test.ts (connection-URL tests for each model name and mode: "manual")
  • prettier, tsc --noEmit
  • Prod, both models, default mode and mode: "manual": Begin + correct final turn; Datadog Session context initialized shows context.mode: manual

🤖 Generated with Claude Code

Adds assemblyai-stt-3-7-realtime (note taking) and
assemblyai-stt-3-7-realtime-agent (voice agents) for the Universal 3.7 launch.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@ccampbell-aai
ccampbell-aai force-pushed the ccampbell/streaming-stt-3-7 branch from 72edaf7 to 3afa7a7 Compare November 16, 2026 16:28
@ccampbell-aai
ccampbell-aai marked this pull request as ready for review November 16, 2026 16:32
The streaming API accepts mode=manual (client ends turns with ForceEndpoint)
on the universal-3 pro models, including assemblyai-stt-3-7-realtime(-agent).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@ccampbell-aai
ccampbell-aai merged commit 92903c1 into main Nov 19, 2026
4 checks passed
@ccampbell-aai
ccampbell-aai deleted the ccampbell/streaming-stt-3-7 branch November 19, 2026 01:14
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants