chore: update OpenAPI spec [sdk-bot]

Author: OpenRouter SDK Bot

.speakeasy/in.openapi.yaml

@@ -20303,6 +20303,7 @@ components:
           - 'top_p'
           - 'max_tokens'
         supports_implicit_caching: true
+        supports_voice_cloning: false
         tag: 'openai'
         throughput_last_30m:
           p50: 45.2
@@ -20404,6 +20405,10 @@ components:
           type: 'array'
         supports_implicit_caching:
           type: 'boolean'
+        supports_voice_cloning:
+          default: false
+          description: 'Whether this TTS endpoint accepts inline reference audio (`input_references`) for stateless voice cloning. Requests carrying reference audio are only routed to endpoints where this is true.'
+          type: 'boolean'
         tag:
           type: 'string'
         throughput_last_30m:
@@ -21836,6 +21841,68 @@ components:
       oneOf:
         - $ref: '#/components/schemas/ContainerAutoEnvironment'
         - $ref: '#/components/schemas/ContainerReferenceEnvironment'
+    SpeechInputReference:
+      description: 'Reference content part for stateless voice cloning'
+      discriminator:
+        mapping:
+          input_audio: '#/components/schemas/SpeechInputReferenceAudio'
+          text: '#/components/schemas/SpeechInputReferenceText'
+        propertyName: 'type'
+      oneOf:
+        - $ref: '#/components/schemas/SpeechInputReferenceAudio'
+        - $ref: '#/components/schemas/SpeechInputReferenceText'
+    SpeechInputReferenceAudio:
+      description: 'Reference audio input for stateless voice cloning'
+      example:
+        input_audio:
+          data: 'data:audio/wav;base64,UklGRuQXDABXQVZF...'
+        type: 'input_audio'
+      properties:
+        input_audio:
+          $ref: '#/components/schemas/SpeechInputReferenceAudioInput'
+        type:
+          enum:
+            - 'input_audio'
+          type: 'string'
+      required:
+        - 'type'
+        - 'input_audio'
+      type: 'object'
+    SpeechInputReferenceAudioInput:
+      description: 'Reference audio input object'
+      properties:
+        data:
+          description: 'Base64-encoded reference audio (optionally a data URI). Supported audio formats are provider-specific. Limited to 20 MiB of base64 (15 MiB of decoded audio).'
+          example: 'data:audio/wav;base64,UklGRuQXDABXQVZF...'
+          maxLength: 20971520
+          minLength: 1
+          type: 'string'
+        format:
+          description: 'Audio format of the reference audio (e.g., wav, mp3). Optional; most providers detect the format from the audio bytes.'
+          example: 'wav'
+          type: 'string'
+      required:
+        - 'data'
+      type: 'object'
+    SpeechInputReferenceText:
+      description: 'Transcript of the accompanying reference audio'
+      example:
+        text: 'I used to rule the world.'
+        type: 'text'
+      properties:
+        text:
+          description: 'Transcript of the accompanying reference audio.'
+          example: 'I used to rule the world.'
+          maxLength: 10000
+          type: 'string'
+        type:
+          enum:
+            - 'text'
+          type: 'string'
+      required:
+        - 'type'
+        - 'text'
+      type: 'object'
     SpeechRequest:
       description: 'Text-to-speech request input'
       example:
@@ -21849,6 +21916,17 @@ components:
           description: 'Text to synthesize'
           example: 'Hello world'
           type: 'string'
+        input_references:
+          description: 'Reference content for stateless voice cloning: one `input_audio` part carrying the voice sample, optionally accompanied by one `text` part with its transcript. Only routed to endpoints that support voice cloning.'
+          example:
+            - input_audio:
+                data: 'data:audio/wav;base64,UklGRuQXDABXQVZF...'
+              type: 'input_audio'
+            - text: 'I used to rule the world.'
+              type: 'text'
+          items:
+            $ref: '#/components/schemas/SpeechInputReference'
+          type: 'array'
         model:
           description: 'TTS model identifier'
           example: 'mistralai/voxtral-mini-tts-2603'
@@ -28169,6 +28247,7 @@ paths:
                       - 'top_p'
                       - 'max_tokens'
                     supports_implicit_caching: true
+                    supports_voice_cloning: false
                     tag: 'openai'
                     throughput_last_30m:
                       p50: 45.2
@@ -28205,6 +28284,7 @@ paths:
                         - 'top_p'
                         - 'max_tokens'
                       supports_implicit_caching: true
+                      supports_voice_cloning: false
                       tag: 'openai'
                       throughput_last_30m:
                         p50: 45.2