@mastra/voice-google 0.13.0-alpha.0 → 0.14.0-alpha.0

package/CHANGELOG.md

@@ -1,5 +1,38 @@
 # @mastra/voice-google
 
+## 0.14.0-alpha.0
+
+### Minor Changes
+
+- Added support for SSML, custom pronunciations, multi-speaker markup, and Gemini-TTS model selection in `GoogleVoice.speak()`. Existing text-only usage is unchanged. ([#19425](https://github.com/mastra-ai/mastra/pull/19425))
+
+  **New `options.input` fields:** pass `ssml`, `markup`, `customPronunciations`, `multiSpeakerMarkup`, or `prompt` (for Gemini-TTS style steering) directly to the Google Cloud TTS API.
+
+  **New `options.voice` fields:** set `modelName` (e.g. `gemini-2.5-flash-preview-tts`) or `multiSpeakerVoiceConfig` alongside the default `name` and `languageCode`.
+
+  ```ts
+  // SSML with custom pronunciations
+  await voice.speak('Give Metacam to the patient.', {
+    input: {
+      ssml: '<speak>Give <phoneme alphabet="ipa" ph="mɛtəˈkæm">Metacam</phoneme>.</speak>',
+    },
+  });
+
+  // Gemini-TTS with prompt-driven styling
+  await voice.speak('Hello!', {
+    voice: { name: 'Kore', modelName: 'gemini-2.5-flash-preview-tts' },
+    input: { prompt: 'Warm, calm tone.' },
+  });
+  ```
+
+- Added Cloud Speech-to-Text v2 support to `GoogleVoice.listen()`. Pass `{ v2: true }` in options to use the v2 API, which supports additional audio formats like AAC-in-MP4 (iOS Safari) via `autoDecodingConfig` or `explicitDecodingConfig`. The v1 path remains the default — no breaking changes. ([#19422](https://github.com/mastra-ai/mastra/pull/19422))
+
+## 0.13.0
+
+### Minor Changes
+
+- Random bump ([#18178](https://github.com/mastra-ai/mastra/pull/18178))
+
 ## 0.13.0-alpha.0
 
 ### Minor Changes

package/dist/_types/@internal_voice/dist/_types/@internal_ai-sdk-v5/dist/index.d.ts

@@ -231,6 +231,7 @@ declare namespace _ai_sdk_provider_utils {
         Validator,
         asSchema,
         asValidator,
+        cancelResponseBody,
         combineHeaders,
         convertAsyncIteratorToReadableStream,
         convertBase64ToUint8Array,
@@ -249,13 +250,16 @@ declare namespace _ai_sdk_provider_utils {
         dynamicTool,
         executeTool,
         extractResponseHeaders,
+        fetchWithValidatedRedirects,
         generateId,
         getErrorMessage,
         getFromApi,
         getRuntimeEnvironmentUserAgent,
         injectJsonInstructionIntoMessages,
         isAbortError,
+        isBrowserRuntime,
         isParsableJson,
+        isSameOrigin,
         isUrlSupported,
         isValidator,
         jsonSchema,
@@ -494,6 +498,20 @@ export declare type CallSettings = {
  */
 export declare type CallWarning = LanguageModelV2CallWarning;
 
+/**
+ * Cancels a response body to release the underlying connection.
+ *
+ * When a fetch Response is rejected without consuming its body (e.g. a failed
+ * status code, an open-redirect rejection, or a Content-Length that exceeds the
+ * size limit), the underlying TCP socket is not returned to the connection pool
+ * and may stay open until the process runs out of file descriptors. Cancelling
+ * the body avoids this leak.
+ *
+ * Errors thrown while cancelling are ignored: the body may already be locked,
+ * disturbed, or absent, none of which should mask the original rejection.
+ */
+declare function cancelResponseBody(response: Response): Promise<void>;
+
 export declare interface ChatInit<UI_MESSAGE extends UIMessage> {
     /**
      * A unique identifier for the chat. If not provided, a random one will be
@@ -956,7 +974,21 @@ export declare type CreateUIMessage<UI_MESSAGE extends UIMessage> = Omit<UI_MESS
     role?: UI_MESSAGE['role'];
 };
 
-export declare function createUIMessageStream<UI_MESSAGE extends UIMessage>({ execute, onError, originalMessages, onFinish, generateId, }: {
+/**
+ * Creates a UI message stream that can be used to send messages to the client.
+ *
+ * @param options.execute - A function that is called with a writer to write UI message chunks to the stream.
+ * @param options.onError - A function that extracts an error message from an error. Defaults to `() => 'An error occurred.'` so server-side error details are not leaked to the client; supply your own to surface richer messages.
+ * @param options.originalMessages - The original messages. If provided, persistence mode is assumed
+ *   and a message ID is provided for the response message.
+ * @param options.onStepFinish - A callback that is called when each step finishes. Useful for persisting intermediate messages.
+ * @param options.onFinish - A callback that is called when the stream finishes.
+ * @param options.generateId - A function that generates a unique ID. Defaults to the built-in ID generator.
+ *
+ * @returns A `ReadableStream` of UI message chunks.
+ */
+export declare function createUIMessageStream<UI_MESSAGE extends UIMessage>({ execute, onError, // prevent leaking server error details to the client by default
+    originalMessages, onFinish, generateId, }: {
     execute: (options: {
         writer: UIMessageStreamWriter<UI_MESSAGE>;
     }) => Promise<void> | void;
@@ -2129,6 +2161,36 @@ declare function extractResponseHeaders(response: Response): {
  */
 declare type FetchFunction = typeof globalThis.fetch;
 
+/**
+ * Fetches a URL while enforcing the SSRF download guard on every hop.
+ *
+ * Redirects are followed manually (`redirect: 'manual'`) so each hop is
+ * validated with {@link validateDownloadUrl} *before* it is requested. Relying
+ * on the default `redirect: 'follow'` would issue the request to a redirect
+ * target (e.g. an internal address) before we ever see its URL, defeating the
+ * SSRF guard.
+ *
+ * A `redirect: 'manual'` request yields an unreadable opaque response in the
+ * browser (and in other spec-compliant fetch implementations), so the redirect
+ * target cannot be validated here. In a real browser this is safe to follow
+ * natively because SSRF is not reachable (fetch is constrained by CORS and
+ * cannot reach a server's internal network or cloud-metadata). On any other
+ * runtime we cannot validate the hop, so we fail closed rather than follow it
+ * blindly and bypass the SSRF guard.
+ *
+ * The returned response is the final (non-redirect) response. The caller is
+ * responsible for checking `response.ok` and reading the body.
+ *
+ * @throws DownloadError if a hop is unsafe, the redirect limit is exceeded, or
+ * a redirect cannot be validated on a non-browser runtime.
+ */
+declare function fetchWithValidatedRedirects({ url, headers, abortSignal, maxRedirects, }: {
+    url: string;
+    headers?: HeadersInit;
+    abortSignal?: AbortSignal;
+    maxRedirects?: number;
+}): Promise<Response>;
+
 /**
  File content part of a prompt. It contains a file.
  */
@@ -2261,7 +2323,7 @@ declare interface GatewayGenerationInfoParams {
     id: string;
 }
 
-declare type GatewayImageModelId = 'bfl/flux-2-flex' | 'bfl/flux-2-klein-4b' | 'bfl/flux-2-klein-9b' | 'bfl/flux-2-max' | 'bfl/flux-2-pro' | 'bfl/flux-kontext-max' | 'bfl/flux-kontext-pro' | 'bfl/flux-pro-1.0-fill' | 'bfl/flux-pro-1.1' | 'bfl/flux-pro-1.1-ultra' | 'bytedance/seedream-4.0' | 'bytedance/seedream-4.5' | 'bytedance/seedream-5.0-lite' | 'google/imagen-4.0-fast-generate-001' | 'google/imagen-4.0-generate-001' | 'google/imagen-4.0-ultra-generate-001' | 'openai/gpt-image-1' | 'openai/gpt-image-1-mini' | 'openai/gpt-image-1.5' | 'openai/gpt-image-2' | 'prodia/flux-fast-schnell' | 'recraft/recraft-v2' | 'recraft/recraft-v3' | 'recraft/recraft-v4' | 'recraft/recraft-v4-pro' | 'xai/grok-imagine-image' | 'xai/grok-imagine-image-pro' | (string & {});
+declare type GatewayImageModelId = 'bfl/flux-2-flex' | 'bfl/flux-2-klein-4b' | 'bfl/flux-2-klein-9b' | 'bfl/flux-2-max' | 'bfl/flux-2-pro' | 'bfl/flux-kontext-max' | 'bfl/flux-kontext-pro' | 'bfl/flux-pro-1.0-fill' | 'bfl/flux-pro-1.1' | 'bfl/flux-pro-1.1-ultra' | 'bytedance/seedream-4.0' | 'bytedance/seedream-4.5' | 'bytedance/seedream-5.0-lite' | 'google/imagen-4.0-fast-generate-001' | 'google/imagen-4.0-generate-001' | 'google/imagen-4.0-ultra-generate-001' | 'openai/gpt-image-1' | 'openai/gpt-image-1-mini' | 'openai/gpt-image-1.5' | 'openai/gpt-image-2' | 'prodia/flux-fast-schnell' | 'quiverai/arrow-1.1' | 'recraft/recraft-v2' | 'recraft/recraft-v3' | 'recraft/recraft-v4' | 'recraft/recraft-v4-pro' | 'recraft/recraft-v4.1' | 'recraft/recraft-v4.1-pro' | 'recraft/recraft-v4.1-utility' | 'recraft/recraft-v4.1-utility-pro' | 'xai/grok-imagine-image' | (string & {});
 
 declare interface GatewayLanguageModelEntry {
     /**
@@ -2312,7 +2374,7 @@ declare interface GatewayLanguageModelEntry {
 
 declare type GatewayLanguageModelSpecification = Pick<LanguageModelV2, 'specificationVersion' | 'provider' | 'modelId'>;
 
-export declare type GatewayModelId = 'alibaba/qwen-3-14b' | 'alibaba/qwen-3-235b' | 'alibaba/qwen-3-30b' | 'alibaba/qwen-3-32b' | 'alibaba/qwen-3.6-max-preview' | 'alibaba/qwen3-235b-a22b-thinking' | 'alibaba/qwen3-coder' | 'alibaba/qwen3-coder-30b-a3b' | 'alibaba/qwen3-coder-next' | 'alibaba/qwen3-coder-plus' | 'alibaba/qwen3-max' | 'alibaba/qwen3-max-preview' | 'alibaba/qwen3-max-thinking' | 'alibaba/qwen3-next-80b-a3b-instruct' | 'alibaba/qwen3-next-80b-a3b-thinking' | 'alibaba/qwen3-vl-235b-a22b-instruct' | 'alibaba/qwen3-vl-instruct' | 'alibaba/qwen3-vl-thinking' | 'alibaba/qwen3.5-flash' | 'alibaba/qwen3.5-plus' | 'alibaba/qwen3.6-27b' | 'alibaba/qwen3.6-plus' | 'amazon/nova-2-lite' | 'amazon/nova-lite' | 'amazon/nova-micro' | 'amazon/nova-pro' | 'anthropic/claude-3-haiku' | 'anthropic/claude-3.5-haiku' | 'anthropic/claude-3.7-sonnet' | 'anthropic/claude-haiku-4.5' | 'anthropic/claude-opus-4' | 'anthropic/claude-opus-4.1' | 'anthropic/claude-opus-4.5' | 'anthropic/claude-opus-4.6' | 'anthropic/claude-opus-4.7' | 'anthropic/claude-sonnet-4' | 'anthropic/claude-sonnet-4.5' | 'anthropic/claude-sonnet-4.6' | 'arcee-ai/trinity-large-preview' | 'arcee-ai/trinity-large-thinking' | 'arcee-ai/trinity-mini' | 'bytedance/seed-1.6' | 'bytedance/seed-1.8' | 'cohere/command-a' | 'deepseek/deepseek-r1' | 'deepseek/deepseek-v3' | 'deepseek/deepseek-v3.1' | 'deepseek/deepseek-v3.1-terminus' | 'deepseek/deepseek-v3.2' | 'deepseek/deepseek-v3.2-thinking' | 'deepseek/deepseek-v4-flash' | 'deepseek/deepseek-v4-pro' | 'google/gemini-2.0-flash' | 'google/gemini-2.0-flash-lite' | 'google/gemini-2.5-flash' | 'google/gemini-2.5-flash-image' | 'google/gemini-2.5-flash-lite' | 'google/gemini-2.5-pro' | 'google/gemini-3-flash' | 'google/gemini-3-pro-image' | 'google/gemini-3-pro-preview' | 'google/gemini-3.1-flash-image-preview' | 'google/gemini-3.1-flash-lite-preview' | 'google/gemini-3.1-pro-preview' | 'google/gemma-4-26b-a4b-it' | 'google/gemma-4-31b-it' | 'inception/mercury-2' | 'inception/mercury-coder-small' | 'interfaze/interfaze-beta' | 'kwaipilot/kat-coder-pro-v1' | 'kwaipilot/kat-coder-pro-v2' | 'meituan/longcat-flash-chat' | 'meituan/longcat-flash-thinking-2601' | 'meta/llama-3.1-70b' | 'meta/llama-3.1-8b' | 'meta/llama-3.2-11b' | 'meta/llama-3.2-1b' | 'meta/llama-3.2-3b' | 'meta/llama-3.2-90b' | 'meta/llama-3.3-70b' | 'meta/llama-4-maverick' | 'meta/llama-4-scout' | 'minimax/minimax-m2' | 'minimax/minimax-m2.1' | 'minimax/minimax-m2.1-lightning' | 'minimax/minimax-m2.5' | 'minimax/minimax-m2.5-highspeed' | 'minimax/minimax-m2.7' | 'minimax/minimax-m2.7-highspeed' | 'mistral/codestral' | 'mistral/devstral-2' | 'mistral/devstral-small' | 'mistral/devstral-small-2' | 'mistral/magistral-medium' | 'mistral/magistral-small' | 'mistral/ministral-14b' | 'mistral/ministral-3b' | 'mistral/ministral-8b' | 'mistral/mistral-large-3' | 'mistral/mistral-medium' | 'mistral/mistral-nemo' | 'mistral/mistral-small' | 'mistral/pixtral-12b' | 'mistral/pixtral-large' | 'moonshotai/kimi-k2' | 'moonshotai/kimi-k2-thinking' | 'moonshotai/kimi-k2-thinking-turbo' | 'moonshotai/kimi-k2-turbo' | 'moonshotai/kimi-k2.5' | 'moonshotai/kimi-k2.6' | 'morph/morph-v3-fast' | 'morph/morph-v3-large' | 'nvidia/nemotron-3-nano-30b-a3b' | 'nvidia/nemotron-3-super-120b-a12b' | 'nvidia/nemotron-nano-12b-v2-vl' | 'nvidia/nemotron-nano-9b-v2' | 'openai/gpt-3.5-turbo' | 'openai/gpt-3.5-turbo-instruct' | 'openai/gpt-4-turbo' | 'openai/gpt-4.1' | 'openai/gpt-4.1-mini' | 'openai/gpt-4.1-nano' | 'openai/gpt-4o' | 'openai/gpt-4o-mini' | 'openai/gpt-4o-mini-search-preview' | 'openai/gpt-5' | 'openai/gpt-5-chat' | 'openai/gpt-5-codex' | 'openai/gpt-5-mini' | 'openai/gpt-5-nano' | 'openai/gpt-5-pro' | 'openai/gpt-5.1-codex' | 'openai/gpt-5.1-codex-max' | 'openai/gpt-5.1-codex-mini' | 'openai/gpt-5.1-instant' | 'openai/gpt-5.1-thinking' | 'openai/gpt-5.2' | 'openai/gpt-5.2-chat' | 'openai/gpt-5.2-codex' | 'openai/gpt-5.2-pro' | 'openai/gpt-5.3-chat' | 'openai/gpt-5.3-codex' | 'openai/gpt-5.4' | 'openai/gpt-5.4-mini' | 'openai/gpt-5.4-nano' | 'openai/gpt-5.4-pro' | 'openai/gpt-5.5' | 'openai/gpt-5.5-pro' | 'openai/gpt-oss-120b' | 'openai/gpt-oss-20b' | 'openai/gpt-oss-safeguard-20b' | 'openai/o1' | 'openai/o3' | 'openai/o3-deep-research' | 'openai/o3-mini' | 'openai/o3-pro' | 'openai/o4-mini' | 'perplexity/sonar' | 'perplexity/sonar-pro' | 'perplexity/sonar-reasoning-pro' | 'xai/grok-3' | 'xai/grok-3-fast' | 'xai/grok-3-mini' | 'xai/grok-3-mini-fast' | 'xai/grok-4' | 'xai/grok-4-fast-non-reasoning' | 'xai/grok-4-fast-reasoning' | 'xai/grok-4.1-fast-non-reasoning' | 'xai/grok-4.1-fast-reasoning' | 'xai/grok-4.20-multi-agent' | 'xai/grok-4.20-multi-agent-beta' | 'xai/grok-4.20-non-reasoning' | 'xai/grok-4.20-non-reasoning-beta' | 'xai/grok-4.20-reasoning' | 'xai/grok-4.20-reasoning-beta' | 'xai/grok-4.3' | 'xai/grok-code-fast-1' | 'xiaomi/mimo-v2-flash' | 'xiaomi/mimo-v2-pro' | 'xiaomi/mimo-v2.5' | 'xiaomi/mimo-v2.5-pro' | 'zai/glm-4.5' | 'zai/glm-4.5-air' | 'zai/glm-4.5v' | 'zai/glm-4.6' | 'zai/glm-4.6v' | 'zai/glm-4.6v-flash' | 'zai/glm-4.7' | 'zai/glm-4.7-flash' | 'zai/glm-4.7-flashx' | 'zai/glm-5' | 'zai/glm-5-turbo' | 'zai/glm-5.1' | 'zai/glm-5v-turbo' | (string & {});
+export declare type GatewayModelId = 'alibaba/qwen-3-14b' | 'alibaba/qwen-3-235b' | 'alibaba/qwen-3-30b' | 'alibaba/qwen-3-32b' | 'alibaba/qwen-3.6-max-preview' | 'alibaba/qwen3-235b-a22b-thinking' | 'alibaba/qwen3-coder' | 'alibaba/qwen3-coder-30b-a3b' | 'alibaba/qwen3-coder-next' | 'alibaba/qwen3-coder-plus' | 'alibaba/qwen3-max' | 'alibaba/qwen3-max-preview' | 'alibaba/qwen3-max-thinking' | 'alibaba/qwen3-next-80b-a3b-instruct' | 'alibaba/qwen3-next-80b-a3b-thinking' | 'alibaba/qwen3-vl-235b-a22b-instruct' | 'alibaba/qwen3-vl-instruct' | 'alibaba/qwen3-vl-thinking' | 'alibaba/qwen3.5-flash' | 'alibaba/qwen3.5-plus' | 'alibaba/qwen3.6-27b' | 'alibaba/qwen3.6-plus' | 'alibaba/qwen3.7-max' | 'alibaba/qwen3.7-plus' | 'amazon/nova-2-lite' | 'amazon/nova-lite' | 'amazon/nova-micro' | 'amazon/nova-pro' | 'anthropic/claude-3-haiku' | 'anthropic/claude-3.5-haiku' | 'anthropic/claude-fable-5' | 'anthropic/claude-haiku-4.5' | 'anthropic/claude-opus-4' | 'anthropic/claude-opus-4.1' | 'anthropic/claude-opus-4.5' | 'anthropic/claude-opus-4.6' | 'anthropic/claude-opus-4.7' | 'anthropic/claude-opus-4.8' | 'anthropic/claude-sonnet-4' | 'anthropic/claude-sonnet-4.5' | 'anthropic/claude-sonnet-4.6' | 'anthropic/claude-sonnet-5' | 'arcee-ai/trinity-large-preview' | 'arcee-ai/trinity-large-thinking' | 'arcee-ai/trinity-mini' | 'bytedance/seed-1.6' | 'bytedance/seed-1.8' | 'cohere/command-a' | 'deepseek/deepseek-r1' | 'deepseek/deepseek-v3' | 'deepseek/deepseek-v3.1' | 'deepseek/deepseek-v3.1-terminus' | 'deepseek/deepseek-v3.2' | 'deepseek/deepseek-v3.2-thinking' | 'deepseek/deepseek-v4-flash' | 'deepseek/deepseek-v4-pro' | 'google/gemini-2.5-flash' | 'google/gemini-2.5-flash-image' | 'google/gemini-2.5-flash-lite' | 'google/gemini-2.5-pro' | 'google/gemini-3-flash' | 'google/gemini-3-pro-image' | 'google/gemini-3-pro-preview' | 'google/gemini-3.1-flash-image' | 'google/gemini-3.1-flash-image-preview' | 'google/gemini-3.1-flash-lite' | 'google/gemini-3.1-flash-lite-image' | 'google/gemini-3.1-flash-lite-preview' | 'google/gemini-3.1-pro-preview' | 'google/gemini-3.5-flash' | 'google/gemma-4-26b-a4b-it' | 'google/gemma-4-31b-it' | 'inception/mercury-2' | 'inception/mercury-coder-small' | 'interfaze/interfaze-beta' | 'kwaipilot/kat-coder-pro-v1' | 'kwaipilot/kat-coder-pro-v2' | 'meituan/longcat-flash-chat' | 'meituan/longcat-flash-thinking-2601' | 'meta/llama-3.1-70b' | 'meta/llama-3.1-8b' | 'meta/llama-3.2-11b' | 'meta/llama-3.2-1b' | 'meta/llama-3.2-3b' | 'meta/llama-3.2-90b' | 'meta/llama-3.3-70b' | 'meta/llama-4-maverick' | 'meta/llama-4-scout' | 'minimax/minimax-m2' | 'minimax/minimax-m2.1' | 'minimax/minimax-m2.1-lightning' | 'minimax/minimax-m2.5' | 'minimax/minimax-m2.5-highspeed' | 'minimax/minimax-m2.7' | 'minimax/minimax-m2.7-highspeed' | 'minimax/minimax-m3' | 'mistral/codestral' | 'mistral/devstral-2' | 'mistral/devstral-small' | 'mistral/devstral-small-2' | 'mistral/magistral-medium' | 'mistral/magistral-small' | 'mistral/ministral-14b' | 'mistral/ministral-3b' | 'mistral/ministral-8b' | 'mistral/mistral-large-3' | 'mistral/mistral-medium' | 'mistral/mistral-medium-3.5' | 'mistral/mistral-nemo' | 'mistral/mistral-small' | 'mistral/pixtral-12b' | 'mistral/pixtral-large' | 'moonshotai/kimi-k2' | 'moonshotai/kimi-k2-thinking' | 'moonshotai/kimi-k2.5' | 'moonshotai/kimi-k2.6' | 'moonshotai/kimi-k2.7-code' | 'moonshotai/kimi-k2.7-code-highspeed' | 'morph/morph-v3-fast' | 'morph/morph-v3-large' | 'nvidia/nemotron-3-nano-30b-a3b' | 'nvidia/nemotron-3-super-120b-a12b' | 'nvidia/nemotron-3-ultra-550b-a55b' | 'nvidia/nemotron-nano-12b-v2-vl' | 'nvidia/nemotron-nano-9b-v2' | 'openai/gpt-3.5-turbo' | 'openai/gpt-3.5-turbo-instruct' | 'openai/gpt-4-turbo' | 'openai/gpt-4.1' | 'openai/gpt-4.1-mini' | 'openai/gpt-4.1-nano' | 'openai/gpt-4o' | 'openai/gpt-4o-mini' | 'openai/gpt-4o-mini-search-preview' | 'openai/gpt-5' | 'openai/gpt-5-chat' | 'openai/gpt-5-codex' | 'openai/gpt-5-mini' | 'openai/gpt-5-nano' | 'openai/gpt-5-pro' | 'openai/gpt-5.1-codex' | 'openai/gpt-5.1-codex-max' | 'openai/gpt-5.1-codex-mini' | 'openai/gpt-5.1-instant' | 'openai/gpt-5.1-thinking' | 'openai/gpt-5.2' | 'openai/gpt-5.2-chat' | 'openai/gpt-5.2-codex' | 'openai/gpt-5.2-pro' | 'openai/gpt-5.3-chat' | 'openai/gpt-5.3-codex' | 'openai/gpt-5.4' | 'openai/gpt-5.4-mini' | 'openai/gpt-5.4-nano' | 'openai/gpt-5.4-pro' | 'openai/gpt-5.5' | 'openai/gpt-5.5-pro' | 'openai/gpt-oss-120b' | 'openai/gpt-oss-20b' | 'openai/gpt-oss-safeguard-20b' | 'openai/o1' | 'openai/o3' | 'openai/o3-deep-research' | 'openai/o3-mini' | 'openai/o3-pro' | 'openai/o4-mini' | 'perplexity/sonar' | 'perplexity/sonar-pro' | 'perplexity/sonar-reasoning-pro' | 'sakana/fugu-ultra' | 'stepfun/step-3.5-flash' | 'stepfun/step-3.7-flash' | 'xai/grok-4.1-fast-non-reasoning' | 'xai/grok-4.1-fast-reasoning' | 'xai/grok-4.20-multi-agent' | 'xai/grok-4.20-multi-agent-beta' | 'xai/grok-4.20-non-reasoning' | 'xai/grok-4.20-non-reasoning-beta' | 'xai/grok-4.20-reasoning' | 'xai/grok-4.20-reasoning-beta' | 'xai/grok-4.3' | 'xai/grok-build-0.1' | 'xiaomi/mimo-v2-flash' | 'xiaomi/mimo-v2-pro' | 'xiaomi/mimo-v2.5' | 'xiaomi/mimo-v2.5-pro' | 'zai/glm-4.5' | 'zai/glm-4.5-air' | 'zai/glm-4.5v' | 'zai/glm-4.6' | 'zai/glm-4.6v' | 'zai/glm-4.6v-flash' | 'zai/glm-4.7' | 'zai/glm-4.7-flash' | 'zai/glm-4.7-flashx' | 'zai/glm-5' | 'zai/glm-5-turbo' | 'zai/glm-5.1' | 'zai/glm-5.2' | 'zai/glm-5.2-fast' | 'zai/glm-5v-turbo' | (string & {});
 
 declare interface GatewayProvider extends ProviderV2 {
     (modelId: GatewayModelId): LanguageModelV2;
@@ -3424,6 +3486,16 @@ export declare class InvalidToolInputError extends AISDKError {
 
 declare function isAbortError(error: unknown): error is Error;
 
+/**
+ * Returns `true` when running in a browser.
+ *
+ * Detection keys on the presence of a global `window`, matching the browser
+ * check used elsewhere in this package (see `getRuntimeEnvironmentUserAgent`)
+ * so the SDK has a single, consistent definition of "browser". Server runtimes
+ * (Node.js, Deno, Bun, edge/workers) do not define `window`.
+ */
+declare function isBrowserRuntime(globalThisAny?: any): boolean;
+
 /**
  * Check if a message part is a data part.
  */
@@ -3456,6 +3528,20 @@ declare function isParsableJson(input: string): boolean;
  */
 export declare function isReasoningUIPart(part: UIMessagePart<UIDataTypes, UITools>): part is ReasoningUIPart;
 
+/**
+ * Returns true when `url` has the same origin (scheme + host + port) as
+ * `baseUrl`.
+ *
+ * Used to decide whether provider credentials may be attached to a request to a
+ * URL taken from a provider response (e.g. a polling or media-download URL).
+ * Credentials must only be sent to the provider's own origin; a response that
+ * names a foreign host (a CDN, or an attacker-controlled host if the response
+ * is tampered with) must not receive the API key.
+ *
+ * Returns false if either value is not a valid absolute URL (fail-closed).
+ */
+declare function isSameOrigin(url: string, baseUrl: string): boolean;
+
 /**
  * Type guard to check if a message part is a text part.
  */
@@ -8714,6 +8800,10 @@ export declare const userModelMessageSchema: z.ZodType<UserModelMessage>;
  * Validates that a URL is safe to download from, blocking private/internal addresses
  * to prevent SSRF attacks.
  *
+ * Note: this performs string/literal-IP checks only. It does not resolve DNS, so a
+ * hostname that resolves to a private address is not blocked here (see callers, which
+ * should additionally constrain egress at the network layer when handling untrusted URLs).
+ *
  * @param url - The URL string to validate.
  * @throws DownloadError if the URL is unsafe.
  */

package/dist/_types/@internal_voice/dist/_types/@internal_core/dist/request-context/index.d.ts

@@ -129,6 +129,16 @@ declare class RequestContext<Values extends Record<string, any> | unknown = unkn
      * a clean JSON-safe dict for cross-context cycles.
      */
     private isSerializable;
+    /**
+     * Custom span serialization to prevent leaking internal state (like auth
+     * tokens stored in the private `registry` Map) into observability spans.
+     *
+     * `deepClean` in `@mastra/observability` calls this method before falling
+     * back to `Object.keys()` — which would walk the runtime-enumerable
+     * `registry` field and serialize its raw Map entries (including any
+     * bearer tokens) into exported spans.
+     */
+    serializeForSpan(): Record<string, unknown>;
     /**
      * Get all values as a typed object for destructuring.
      * Returns Record<string, any> when untyped, or the Values type when typed.

package/dist/docs/SKILL.md

@@ -3,7 +3,7 @@ name: mastra-voice-google
 description: Documentation for @mastra/voice-google. Use when working with @mastra/voice-google APIs, configuration, or implementation.
 metadata:
   package: "@mastra/voice-google"
-  version: "0.13.0-alpha.0"
+  version: "0.14.0-alpha.0"
 ---
 
 ## When to use
@@ -16,7 +16,6 @@ Read the individual reference documents for detailed explanations and code examp
 
 ### Docs
 
-- [Voice](references/docs-agents-adding-voice.md) - Learn how to add voice capabilities to your Mastra agents for text-to-speech and speech-to-text interactions.
 - [Voice in Mastra](references/docs-voice-overview.md) - Overview of voice capabilities in Mastra, including text-to-speech, speech-to-text, and real-time speech-to-speech interactions.
 
 ### Reference

package/dist/docs/assets/SOURCE_MAP.json

@@ -1,5 +1,5 @@
 {
-  "version": "0.13.0-alpha.0",
+  "version": "0.14.0-alpha.0",
   "package": "@mastra/voice-google",
   "exports": {},
   "modules": {}

package/dist/docs/references/docs-voice-overview.md

@@ -1,10 +1,12 @@
+> Discover all available pages from the documentation index: https://mastra.ai/llms.txt
+
 # Voice in Mastra
 
 Mastra's Voice system provides a unified interface for voice interactions, enabling text-to-speech (TTS), speech-to-text (STT), and real-time speech-to-speech (STS) capabilities in your applications.
 
-## Adding voice to agents
+## Add voice to agents
 
-To learn how to integrate voice capabilities into your agents, check out the [Adding Voice to Agents](https://mastra.ai/docs/agents/adding-voice) documentation. This section covers how to use both single and multiple voice providers, as well as real-time interactions.
+Pass a voice provider to an agent with the `voice` property. The same property supports text-to-speech (TTS), speech-to-text (STT), and real-time speech-to-speech (STS), depending on the provider you configure.
 
 ```typescript
 import { Agent } from '@mastra/core/agent'
@@ -1137,7 +1139,7 @@ const voiceAgent = new Agent({
 })
 ```
 
-### Using Multiple Voice Providers
+### Using multiple voice providers
 
 This example demonstrates how to create and use two different voice providers in Mastra: OpenAI for speech-to-text (STT) and PlayAI for text-to-speech (TTS).
 

package/dist/docs/references/reference-voice-google.md

@@ -1,3 +1,5 @@
+> Discover all available pages from the documentation index: https://mastra.ai/llms.txt
+
 # Google
 
 The Google Voice implementation in Mastra provides both text-to-speech (TTS) and speech-to-text (STT) capabilities using Google Cloud services. It supports multiple voices, languages, advanced audio configuration options, and both standard API key authentication and Vertex AI mode for enterprise deployments.
@@ -10,7 +12,7 @@ import { GoogleVoice } from '@mastra/voice-google'
 // Initialize with default configuration (uses GOOGLE_API_KEY environment variable)
 const voice = new GoogleVoice()
 
-// Text-to-Speech
+// Text-to-Speech (plain text)
 const audioStream = await voice.speak('Hello, world!', {
   languageCode: 'en-US',
   audioConfig: {
@@ -18,6 +20,19 @@ const audioStream = await voice.speak('Hello, world!', {
   },
 })
 
+// Text-to-Speech with SSML
+const ssmlStream = await voice.speak('ignored', {
+  input: {
+    ssml: '<speak>Take <say-as interpret-as="unit">5 mg</say-as> daily.</speak>',
+  },
+})
+
+// Text-to-Speech with Gemini-TTS model
+const geminiStream = await voice.speak('Hello from Gemini TTS!', {
+  voice: { name: 'Kore', modelName: 'gemini-2.5-flash-preview-tts' },
+  input: { prompt: 'Warm, calm tone.' },
+})
+
 // Speech-to-Text
 const transcript = await voice.listen(audioStream, {
   config: {
@@ -66,25 +81,52 @@ Converts text to speech using Google Cloud Text-to-Speech service.
 
 **options** (`object`): Speech synthesis options
 
-**options.speaker** (`string`): Voice ID to use for this request
+**options.speaker** (`string`): Voice ID to use for this request.
 
-**options.languageCode** (`string`): Language code for the voice (e.g., 'en-US'). Defaults to the language code from the speaker ID or 'en-US'
+**options.languageCode** (`string`): Language code for the voice (e.g., 'en-US'). Defaults to the language code derived from the speaker ID, or 'en-US'.
 
-**options.audioConfig** (`ISynthesizeSpeechRequest['audioConfig']`): Audio configuration options from Google Cloud Text-to-Speech API
+**options.input** (`ISynthesizeSpeechRequest['input']`): Rich input object passed through to the Google Cloud TTS API. Supports ssml, markup, prompt (Gemini-TTS style steering), customPronunciations, and multiSpeakerMarkup. When provided without text, ssml, markup, or multiSpeakerMarkup, the positional input argument is used as the text field automatically.
+
+**options.voice** (`ISynthesizeSpeechRequest['voice']`): Voice configuration merged on top of defaults (name and languageCode). Supports modelName (e.g., 'gemini-2.5-flash-preview-tts') and multiSpeakerVoiceConfig.
+
+**options.audioConfig** (`ISynthesizeSpeechRequest['audioConfig']`): Audio configuration options from Google Cloud Text-to-Speech API.
 
 Returns: `Promise<NodeJS.ReadableStream>`
 
 ### `listen()`
 
-Converts speech to text using Google Cloud Speech-to-Text service.
+Converts speech to text using Google Cloud Speech-to-Text service. Supports both v1 (default) and v2 APIs. The v2 API adds support for AAC-in-MP4 audio (iOS Safari) via auto-decoding.
+
+#### v1 (default)
+
+**audioStream** (`NodeJS.ReadableStream`): Audio stream to transcribe
+
+**options** (`GoogleListenOptionsV1`): v1 recognition options
+
+**options.config** (`IRecognitionConfig`): v1 recognition configuration from Google Cloud Speech-to-Text API
+
+#### v2
+
+Pass `v2: true` to use the Cloud Speech-to-Text v2 API, which supports additional audio formats like AAC-in-MP4 (iOS Safari).
+
+```typescript
+const transcript = await voice.listen(iosSafariAacStream, {
+  v2: true,
+  config: {
+    autoDecodingConfig: {},
+  },
+})
+```
 
 **audioStream** (`NodeJS.ReadableStream`): Audio stream to transcribe
 
-**options** (`object`): Recognition options
+**options** (`GoogleListenOptionsV2`): v2 recognition options
+
+**options.v2** (`true`): Enables the v2 API path
 
-**options.stream** (`boolean`): Whether to use streaming recognition
+**options.config** (`v2.IRecognitionConfig`): v2 recognition configuration. Defaults to auto-decoding with languageCodes: \['en-US'] and model: 'long'. Set autoDecodingConfig: {} to auto-detect the audio format, or use explicitDecodingConfig to specify an encoding like MP4\_AAC, M4A\_AAC, or MOV\_AAC.
 
-**options.config** (`IRecognitionConfig`): Recognition configuration from Google Cloud Speech-to-Text API
+**options.recognizer** (`string`): v2 recognizer resource path. Defaults to projects/{project}/locations/global/recognizers/\_ where {project} is resolved from the constructor project option, GOOGLE\_CLOUD\_PROJECT, or the client's default project.
 
 Returns: `Promise<string>`
 

package/dist/index.cjs

@@ -308,6 +308,8 @@ var DEFAULT_VOICE = "en-US-Casual-K";
 var GoogleVoice = class extends MastraVoice {
   ttsClient;
   speechClient;
+  speechClientV2;
+  speechOptionsV2;
   vertexAI;
   project;
   location;
@@ -357,6 +359,13 @@ var GoogleVoice = class extends MastraVoice {
     const speechOptions = buildAuthOptions(listeningAuthConfig, { vertexAI});
     this.ttsClient = new textToSpeech.TextToSpeechClient(ttsOptions);
     this.speechClient = new speech.SpeechClient(speechOptions);
+    this.speechOptionsV2 = speechOptions;
+  }
+  getV2SpeechClient() {
+    if (!this.speechClientV2) {
+      this.speechClientV2 = new speech.v2.SpeechClient(this.speechOptionsV2);
+    }
+    return this.speechClientV2;
   }
   /**
    * Check if Vertex AI mode is enabled
@@ -402,21 +411,30 @@ var GoogleVoice = class extends MastraVoice {
     return Buffer.concat(chunks).toString("utf-8");
   }
   /**
-   * Converts text to speech
+   * Converts text to speech.
+   *
+   * When `input` is a string or stream, builds a text-only request (existing behaviour).
+   * Pass `options.input` to send richer proto fields (ssml, markup, prompt,
+   * customPronunciations, multiSpeakerMarkup) and `options.voice` for fields
+   * like modelName or multiSpeakerVoiceConfig.
+   *
    * @param {string | NodeJS.ReadableStream} input - Text or stream to convert to speech
-   * @param {Object} [options] - Speech synthesis options
-   * @param {string} [options.speaker] - Voice ID to use
-   * @param {string} [options.languageCode] - Language code for the voice
-   * @param {TextToSpeechTypes.cloud.texttospeech.v1.ISynthesizeSpeechRequest['audioConfig']} [options.audioConfig] - Audio configuration options
-   * @returns {Promise<NodeJS.ReadableStream>} Stream of synthesized audio. Default encoding is LINEAR16.
+   * @param {GoogleSpeakOptions} [options] - Speech synthesis options
+   * @returns {Promise<NodeJS.ReadableStream>} Stream of synthesised audio. Default encoding is LINEAR16.
    */
   async speak(input, options) {
-    const text = typeof input === "string" ? input : await this.streamToString(input);
+    const defaultVoiceName = options?.speaker || this.speaker;
+    const defaultLanguageCode = options?.languageCode || defaultVoiceName?.split("-").slice(0, 2).join("-") || "en-US";
+    const requestInput = options?.input ? { ...options.input } : { text: typeof input === "string" ? input : await this.streamToString(input) };
+    if (options?.input && !options.input.text && !options.input.ssml && !options.input.markup && !options.input.multiSpeakerMarkup) {
+      requestInput.text = typeof input === "string" ? input : await this.streamToString(input);
+    }
     const request = {
-      input: { text },
+      input: requestInput,
       voice: {
-        name: options?.speaker || this.speaker,
-        languageCode: options?.languageCode || options?.speaker?.split("-").slice(0, 2).join("-") || "en-US"
+        name: defaultVoiceName,
+        languageCode: defaultLanguageCode,
+        ...options?.voice
       },
       audioConfig: options?.audioConfig || { audioEncoding: "LINEAR16" }
     };
@@ -440,10 +458,14 @@ var GoogleVoice = class extends MastraVoice {
     return { enabled: true };
   }
   /**
-   * Converts speech to text
+   * Converts speech to text using Cloud Speech-to-Text v1 or v2.
+   *
+   * Pass `{ v2: true }` in options to use the v2 API, which supports additional
+   * audio formats like AAC-in-MP4 (iOS Safari) via `autoDecodingConfig` or
+   * `explicitDecodingConfig`. The v1 path remains the default.
+   *
    * @param {NodeJS.ReadableStream} audioStream - Audio stream to transcribe. Default encoding is LINEAR16.
-   * @param {Object} [options] - Recognition options
-   * @param {SpeechTypes.cloud.speech.v1.IRecognitionConfig} [options.config] - Recognition configuration
+   * @param {GoogleListenOptions} [options] - Recognition options
    * @returns {Promise<string>} Transcribed text
    */
   async listen(audioStream, options) {
@@ -456,7 +478,13 @@ var GoogleVoice = class extends MastraVoice {
       }
     }
     const buffer = Buffer.concat(chunks);
-    let request = {
+    if (options && "v2" in options && options.v2) {
+      return this.recognizeV2(buffer, options);
+    }
+    return this.recognizeV1(buffer, options);
+  }
+  async recognizeV1(buffer, options) {
+    const request = {
       config: {
         encoding: "LINEAR16",
         languageCode: "en-US",
@@ -467,10 +495,38 @@ var GoogleVoice = class extends MastraVoice {
       }
     };
     const [response] = await this.speechClient.recognize(request);
-    if (!response.results || response.results.length === 0) {
+    return this.extractTranscription(response?.results);
+  }
+  async recognizeV2(buffer, options) {
+    const config = { ...options.config };
+    if (!config.autoDecodingConfig && !config.explicitDecodingConfig) {
+      config.autoDecodingConfig = {};
+    }
+    if (!config.languageCodes || config.languageCodes.length === 0) {
+      config.languageCodes = ["en-US"];
+    }
+    if (!config.model) {
+      config.model = "long";
+    }
+    let recognizer = options.recognizer;
+    if (!recognizer) {
+      const project = this.project || await this.getV2SpeechClient().getProjectId();
+      recognizer = `projects/${project}/locations/global/recognizers/_`;
+    }
+    const request = {
+      recognizer,
+      config,
+      content: buffer
+    };
+    const client = this.getV2SpeechClient();
+    const [response] = await client.recognize(request);
+    return this.extractTranscription(response?.results);
+  }
+  extractTranscription(results) {
+    if (!results || results.length === 0) {
       throw new Error("No transcription results returned");
     }
-    const transcription = response.results.map((result) => {
+    const transcription = results.map((result) => {
       if (!result.alternatives || result.alternatives.length === 0) {
         return "";
       }