v0.55.0 v0.55.1

Name: TTS-cacher
Author: kazuto-07

medium risk

kazuto-07/TTS-cacherOct 2, 2026 → Oct 2, 20260 days between releases

What's New

Defaults to Deepgram Aura TTS and fixes pause-mode audio regression

New Features

Deepgram Aura TTS integration for the bot example
New BOT_TTS_MODEL configuration setting for OpenAI TTS

Improvements

Significantly faster time-to-first audio with Deepgram Aura (0.52s median)
Reduced latency between caller and bot audio (1.95s to 1.01s)
Improved PSTN call start times

Bug Fixes

Fixed regression where pause-mode barge-ins dropped bot audio
Resolves audio dropping during pause arbitration resolution

Upgrade Notes

The `pipecat-bot-py` example now defaults to Deepgram Aura TTS, resulting in faster start times. To use OpenAI TTS, explicitly set `BOT_TTS_PROVIDER=openai` and configure `BOT_TTS_MODEL`. This change may affect latency characteristics depending on your network and provider.

The Two Releases

v0.55.0

Oct 2, 2026· 20 days ago

Introduces a new Pipecat transport SDK and updates the Sightglass UI security dependencies.

Full release notes

v0.55.1

Oct 2, 2026· 20 days ago

Improves TTS performance by defaulting to Deepgram Aura and fixes a regression where pause-mode barge-ins could drop bot audio.

Full release notes