mirror of
https://github.com/openclaw/openclaw.git
synced 2026-10-03 01:29:56 +00:00
* refactor(providers): deslop small provider plugins * test(providers): align cleanup fixtures and assertion allowances * test(exa): reuse guarded request body in wire fixture |
||
|---|---|---|
| .. | ||
| assets | ||
| audio-flux.live.test.ts | ||
| audio-flux.test.ts | ||
| audio-flux.ts | ||
| audio.live.test.ts | ||
| audio.test.ts | ||
| audio.ts | ||
| capability-catalog.ts | ||
| index.ts | ||
| media-understanding-provider.ts | ||
| openclaw.plugin.json | ||
| package.json | ||
| README.md | ||
| realtime-transcription-provider-factory.ts | ||
| realtime-transcription-provider.cold.test.ts | ||
| realtime-transcription-provider.test.ts | ||
| tsconfig.json | ||
Deepgram
Transcribe recorded audio and live speech with Deepgram. The plugin supplies audio understanding for incoming attachments and realtime transcription for supported voice integrations.
Get started
Provide DEEPGRAM_API_KEY in the Gateway's environment. For recorded audio,
select deepgram in tools.media.models with the audio capability and keep
tools.media.audio.enabled enabled. Send a voice note through a connected
channel to use the transcript in the conversation.
Realtime voice integrations have separate transcription settings. This plugin transcribes speech; it does not provide speech synthesis.
See the Deepgram guide and audio guide for configuration and model selection.