feat: Support safety settings for Gemini voices and deprecate relax_safety_filters

Google APIs · copybara-github · commit a481d8bc0fee · 2026-03-31T15:15:13.000-07:00
feat: Support `enable_textnorm` for Gemini voices.
feat: Mark `advanced_voice_options` as optional.
docs: A comment for field `relax_safety_filters` in message `.google.cloud.texttospeech.v1beta1.AdvancedVoiceOptions` is changed
docs: A comment for field `advanced_voice_options` in message `.google.cloud.texttospeech.v1beta1.SynthesizeSpeechRequest` is changed
docs: A comment for field `markup` in message `.google.cloud.texttospeech.v1beta1.SynthesisInput` is changed
docs: A comment for field `markup` in message `.google.cloud.texttospeech.v1beta1.StreamingSynthesisInput` is changed

PiperOrigin-RevId: 892460024
diff --git a/google/cloud/texttospeech/v1/cloud_tts.proto b/google/cloud/texttospeech/v1/cloud_tts.proto
@@ -1,4 +1,4 @@
-// Copyright 2025 Google LLC
+// Copyright 2026 Google LLC
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -161,18 +161,85 @@ message Voice {
 
 // Used for advanced voice options.
 message AdvancedVoiceOptions {
+  // Safety setting for a single harm category.
+  message SafetySetting {
+    // The harm category to apply the safety setting to.
+    HarmCategory category = 1;
+
+    // The harm block threshold for the safety setting.
+    HarmBlockThreshold threshold = 2;
+  }
+
+  // Safety settings for the request.
+  message SafetySettings {
+    // The safety settings for the request.
+    repeated SafetySetting settings = 1;
+  }
+
+  // Harm categories that will block the content.
+  enum HarmCategory {
+    // Default value. This value is unused.
+    HARM_CATEGORY_UNSPECIFIED = 0;
+
+    // Content that promotes violence or incites hatred against individuals or
+    // groups based on certain attributes.
+    HARM_CATEGORY_HATE_SPEECH = 1;
+
+    // Content that promotes, facilitates, or enables dangerous activities.
+    HARM_CATEGORY_DANGEROUS_CONTENT = 2;
+
+    // Abusive, threatening, or content intended to bully, torment, or ridicule.
+    HARM_CATEGORY_HARASSMENT = 3;
+
+    // Content that contains sexually explicit material.
+    HARM_CATEGORY_SEXUALLY_EXPLICIT = 4;
+  }
+
+  // Harm block thresholds for the safety settings.
+  enum HarmBlockThreshold {
+    // The harm block threshold is unspecified.
+    HARM_BLOCK_THRESHOLD_UNSPECIFIED = 0;
+
+    // Block content with a low harm probability or higher.
+    BLOCK_LOW_AND_ABOVE = 1;
+
+    // Block content with a medium harm probability or higher.
+    BLOCK_MEDIUM_AND_ABOVE = 2;
+
+    // Block content with a high harm probability.
+    BLOCK_ONLY_HIGH = 3;
+
+    // Do not block any content, regardless of its harm probability.
+    BLOCK_NONE = 4;
+
+    // Turn off the safety filter entirely.
+    OFF = 5;
+  }
+
   // Only for Journey voices. If false, the synthesis is context aware
   // and has a higher latency.
   optional bool low_latency_journey_synthesis = 1;
 
-  // Optional. Input only. If true, relaxes safety filters for Gemini TTS. Only
-  // supported for accounts linked to Invoiced (Offline) Cloud billing accounts.
-  // Otherwise, will return result
-  // [google.rpc.Code.INVALID_ARGUMENT][google.rpc.Code.INVALID_ARGUMENT].
+  // Optional. Input only. Deprecated, use safety_settings instead.
+  // If true, relaxes safety filters for Gemini TTS.
   bool relax_safety_filters = 8 [
+    deprecated = true,
     (google.api.field_behavior) = INPUT_ONLY,
     (google.api.field_behavior) = OPTIONAL
   ];
+
+  // Optional. Input only. This applies to Gemini TTS only. If set, the category
+  // specified in the safety setting will be blocked if the harm probability is
+  // above the threshold. Otherwise, the safety filter will be disabled by
+  // default.
+  SafetySettings safety_settings = 9 [
+    (google.api.field_behavior) = INPUT_ONLY,
+    (google.api.field_behavior) = OPTIONAL
+  ];
+
+  // Optional. If true, textnorm will be applied to text input. This feature is
+  // enabled by default. Only applies for Gemini TTS.
+  optional bool enable_textnorm = 2 [(google.api.field_behavior) = OPTIONAL];
 }
 
 // The top-level message sent by the client for the `SynthesizeSpeech` method.
@@ -186,8 +253,9 @@ message SynthesizeSpeechRequest {
   // Required. The configuration of the synthesized audio.
   AudioConfig audio_config = 3 [(google.api.field_behavior) = REQUIRED];
 
-  // Advanced voice options.
-  optional AdvancedVoiceOptions advanced_voice_options = 8;
+  // Optional. Advanced voice options.
+  optional AdvancedVoiceOptions advanced_voice_options = 8
+      [(google.api.field_behavior) = OPTIONAL];
 }
 
 // Pronunciation customization for a phrase.
@@ -296,8 +364,8 @@ message SynthesisInput {
     // The raw text to be synthesized.
     string text = 1;
 
-    // Markup for HD voices specifically. This field may not be used with any
-    // other voices.
+    // Markup for Chirp 3: HD voices specifically. This field may not be used
+    // with any other voices.
     string markup = 5;
 
     // The SSML document to be synthesized. The SSML document must be valid
@@ -522,6 +590,10 @@ message StreamingSynthesizeConfig {
   // be inside a phoneme tag.
   CustomPronunciations custom_pronunciations = 5
       [(google.api.field_behavior) = OPTIONAL];
+
+  // Optional. Advanced voice options.
+  optional AdvancedVoiceOptions advanced_voice_options = 7
+      [(google.api.field_behavior) = OPTIONAL];
 }
 
 // Input to be synthesized.
@@ -532,8 +604,8 @@ message StreamingSynthesisInput {
     // in the output audio.
     string text = 1;
 
-    // Markup for HD voices specifically. This field may not be used with any
-    // other voices.
+    // Markup for Chirp 3: HD voices specifically. This field may not be used
+    // with any other voices.
     string markup = 5;
 
     // Multi-speaker markup for Gemini TTS. This field may not
diff --git a/google/cloud/texttospeech/v1/cloud_tts_lrs.proto b/google/cloud/texttospeech/v1/cloud_tts_lrs.proto
@@ -1,4 +1,4 @@
-// Copyright 2025 Google LLC
+// Copyright 2026 Google LLC
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// Copyright 2025 Google LLC`
	`1`	`+// Copyright 2026 Google LLC`
`2`	`2`	`//`
`3`	`3`	`// Licensed under the Apache License, Version 2.0 (the "License");`
`4`	`4`	`// you may not use this file except in compliance with the License.`