@@ -132,6 +132,7 @@ Choose your preferred auth method and follow the setup steps.
132132| Image generation | Yes |
133133| Music generation | Yes |
134134| Text-to-speech | Yes |
135+| Realtime voice | Yes (Google Live API) |
135136| Image understanding | Yes |
136137| Audio transcription | Yes |
137138| Video understanding | Yes |
@@ -281,6 +282,63 @@ A Google Cloud Console API key restricted to the Gemini API is valid for this
281282provider. This is not the separate Cloud Text-to-Speech API path.
282283</Note>
283284285+## Realtime voice
286+287+The bundled `google` plugin registers a realtime voice provider backed by the
288+Gemini Live API for backend audio bridges such as Voice Call and Google Meet.
289+290+| Setting | Config path | Default |
291+| --------------------- | ------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
292+| Model | `plugins.entries.voice-call.config.realtime.providers.google.model` | `gemini-2.5-flash-native-audio-preview-12-2025` |
293+| Voice | `...google.voice` | `Kore` |
294+| Temperature | `...google.temperature` | (unset) |
295+| VAD start sensitivity | `...google.startSensitivity` | (unset) |
296+| VAD end sensitivity | `...google.endSensitivity` | (unset) |
297+| Silence duration | `...google.silenceDurationMs` | (unset) |
298+| API key | `...google.apiKey` | Falls back to `models.providers.google.apiKey`, `GEMINI_API_KEY`, or `GOOGLE_API_KEY` |
299+300+Example Voice Call realtime config:
301+302+```json5
303+{
304+ plugins: {
305+ entries: {
306+"voice-call": {
307+ enabled: true,
308+ config: {
309+ realtime: {
310+ enabled: true,
311+ provider: "google",
312+ providers: {
313+ google: {
314+ model: "gemini-2.5-flash-native-audio-preview-12-2025",
315+ voice: "Kore",
316+ },
317+ },
318+ },
319+ },
320+ },
321+ },
322+ },
323+}
324+```
325+326+<Note>
327+Google Live API uses bidirectional audio and function calling over a WebSocket.
328+OpenClaw adapts telephony/Meet bridge audio to Gemini's PCM Live API stream and
329+keeps tool calls on the shared realtime voice contract. Leave `temperature`
330+unset unless you need sampling changes; OpenClaw omits non-positive values
331+because Google Live can return transcripts without audio for `temperature: 0`.
332+Gemini API transcription is enabled without `languageCodes`; the current Google
333+SDK rejects language-code hints on this API path.
334+</Note>
335+336+<Note>
337+Control UI Talk browser sessions still require a realtime voice provider with a
338+browser WebRTC session implementation. Today that path is OpenAI Realtime; the
339+Google provider is for backend realtime bridges.
340+</Note>
341+284342## Advanced configuration
285343286344<AccordionGroup>
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。