Skip to content

fix: report speech end times across other stt plugins - #7988

Closed
chenghao-mou wants to merge 1 commit into
chenghao/fix/AGT-3678-stt-eos-timestampsfrom
chenghao/fix/AGT-3678-other-stt-eos-timestamps
Closed

chenghao-mou wants to merge 1 commit into
chenghao/fix/AGT-3678-stt-eos-timestampsfrom
chenghao/fix/AGT-3678-other-stt-eos-timestamps

Conversation

@chenghao-mou

Copy link
Copy Markdown
Member

Carry available speech-end timestamps through EOS events across the remaining STT plugins and hosted inference. Reconnects reset provider clocks, while queued RTZR audio and SLNG replay retain their original timing. Providers without verified speech-end timing remain untimed.

Stacked on #7796. Addresses AGT-3678 (Linear). Extends @AndrazP's EOS diagnosis in #7651 and @longcw's reconnect review.

Initial prompt and agent context

Model: GPT-6 Astra

Can you investigate 3678? I believe this is a STT plugin reporting gap.

Can you create a PR to address the EOS speech end time reporting gap?

yes, we need to fix all of them. and verify them one by one please.

should we move other working plugin changes to a different PR? as only 11labs is the one having issues.

Carry provider speech-end timing through endpoint events and re-anchor
provider clocks after internal reconnects. Preserve queued and replayed
audio timing without assigning speech-end times to untimed providers.

Addresses AGT-3678

Co-authored-by: Andraž Pajtler <1963904+AndrazP@users.noreply.github.com>
Co-authored-by: Long Chen <longch1024@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

plugins review_effort:high Needs broad context and substantial validation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant