sstcaiteam/open-webui Docker 镜像 - 轩辕镜像 | Docker 镜像高效稳定拉取服务

New Feature: Speaker Diarization for Enhanced Transcriptions

We're excited to announce a powerful new enhancement to Open WebUI's Speech-to-Text (STT) capabilities: Speaker Diarization (who spoke when), powered by pyannote/speaker-diarization. This feature intelligently identifies and differentiates individual speakers within your audio files, providing much clearer and more organized transcriptions, especially for meetings, interviews, or multi-person discussions.

How Speaker Diarization Benefits You:

• Clearer Transcripts: Easily follow who said what, making it simpler to review conversations.
• Improved Readability: Say goodbye to long blocks of undifferentiated text. Transcripts are automatically sectioned by speaker.
• Efficient Information Extraction: Quickly pinpoint contributions from specific participants in an audio recording.

Getting Started with Speaker Diarization:

Using this new feature is straightforward:

1. Run Open WebUI with docker:
   docker run -d -p 3000:8080 --gpus all -v open-webui:/app/backend/data --name open-webui-diar sstcaiteam/open-webui:v0.6.34-diar-cuda

2. Enable the Feature:

   • Navigate to your Open WebUI Admin Panel.
   • Click on Settings, then select the Audio page.
   • Toggle the Enable Diarization option to "On." and fill the Hugging Face Token.

3. Upload Your Audio:

   • Simply drag and drop your meeting audio file (supported formats include .mp3 and .wav) directly into the Open WebUI interface.

4. View Your Enhanced Transcript:

   • Open WebUI will process the audio. Once complete, you'll see the transcription in your chat output, now clearly marked with speaker labels.

Example Output:

00:00:00-00:00:10 SPEAKER_00: Hello everyone, thanks for joining today's meeting.
00:00:13-00:00:16 SPEAKER_02: Hi! Happy to be here.
00:00:18-00:00:24 SPEAKER_00: We'll be discussing the Q3 marketing strategy.
00:00:26-00:00:31 SPEAKER_03: I've prepared some initial thoughts on that.

We believe this feature will significantly enhance your experience with Open WebUI's transcription services.