@cartesia/cartesia-js 2.1.2 → 2.1.4

package/reference.md

@@ -246,6 +246,99 @@ await client.datasets.uploadFile(fs.createReadStream("/path/to/your/file"), "str
 </dl>
 </details>
 
+## Infill
+
+<details><summary><code>client.infill.<a href="/src/api/resources/infill/client/Client.ts">bytes</a>(leftAudio, rightAudio, { ...params }) -> stream.Readable</code></summary>
+<dl>
+<dd>
+
+#### 📝 Description
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+Generate audio that smoothly connects two existing audio segments. This is useful for inserting new speech between existing speech segments while maintaining natural transitions.
+
+Only the `sonic-preview` model is supported for infill at this time.
+
+At least one of `left_audio` or `right_audio` must be provided.
+
+</dd>
+</dl>
+</dd>
+</dl>
+
+#### 🔌 Usage
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+```typescript
+await client.infill.bytes(fs.createReadStream("/path/to/your/file"), fs.createReadStream("/path/to/your/file"), {
+    modelId: "sonic-preview",
+    language: "en",
+    transcript: "middle segment",
+    voiceId: "694f9389-aac1-45b6-b726-9d9369183238",
+    outputFormatContainer: "mp3",
+    outputFormatSampleRate: 44100,
+    outputFormatBitRate: 128000,
+});
+```
+
+</dd>
+</dl>
+</dd>
+</dl>
+
+#### ⚙️ Parameters
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+**leftAudio:** `File | fs.ReadStream | Blob`
+
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**rightAudio:** `File | fs.ReadStream | Blob`
+
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**request:** `Cartesia.InfillBytesRequest`
+
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**requestOptions:** `Infill.RequestOptions`
+
+</dd>
+</dl>
+</dd>
+</dl>
+
+</dd>
+</dl>
+</details>
+
 ## Tts
 
 <details><summary><code>client.tts.<a href="/src/api/resources/tts/client/Client.ts">bytes</a>({ ...params }) -> stream.Readable</code></summary>
@@ -584,7 +677,6 @@ await client.voices.create({
         1, 1, 1, 1, 1, 1, 1,
     ],
     language: "en",
-    baseVoiceId: "string",
 });
 ```
 

package/serialization/resources/tts/types/Emotion.d.ts

@@ -6,5 +6,5 @@ import * as Cartesia from "../../../../api/index";
 import * as core from "../../../../core";
 export declare const Emotion: core.serialization.Schema<serializers.Emotion.Raw, Cartesia.Emotion>;
 export declare namespace Emotion {
-    type Raw = "anger:lowest" | "anger:low" | "anger:high" | "anger:highest" | "positivity:lowest" | "positivity:low" | "positivity:high" | "positivity:highest" | "surprise:lowest" | "surprise:high" | "surprise:highest" | "sadness:lowest" | "sadness:low" | "curiosity:low" | "curiosity:high" | "curiosity:highest";
+    type Raw = "anger:lowest" | "anger:low" | "anger:high" | "anger:highest" | "positivity:lowest" | "positivity:low" | "positivity:high" | "positivity:highest" | "surprise:lowest" | "surprise:low" | "surprise:high" | "surprise:highest" | "sadness:lowest" | "sadness:low" | "sadness:high" | "sadness:highest" | "curiosity:lowest" | "curiosity:low" | "curiosity:high" | "curiosity:highest";
 }

package/serialization/resources/tts/types/Emotion.js

@@ -38,10 +38,14 @@ exports.Emotion = core.serialization.enum_([
     "positivity:high",
     "positivity:highest",
     "surprise:lowest",
+    "surprise:low",
     "surprise:high",
     "surprise:highest",
     "sadness:lowest",
     "sadness:low",
+    "sadness:high",
+    "sadness:highest",
+    "curiosity:lowest",
     "curiosity:low",
     "curiosity:high",
     "curiosity:highest",

package/serialization/resources/tts/types/GenerationRequest.d.ts

@@ -20,5 +20,6 @@ export declare namespace GenerationRequest {
         context_id: ContextId.Raw;
         continue?: boolean | null;
         add_timestamps?: boolean | null;
+        add_phoneme_timestamps?: boolean | null;
     }
 }

package/serialization/resources/tts/types/GenerationRequest.js

@@ -42,4 +42,5 @@ exports.GenerationRequest = core.serialization.object({
     contextId: core.serialization.property("context_id", ContextId_1.ContextId),
     continue: core.serialization.boolean().optional(),
     addTimestamps: core.serialization.property("add_timestamps", core.serialization.boolean().optional()),
+    addPhonemeTimestamps: core.serialization.property("add_phoneme_timestamps", core.serialization.boolean().optional()),
 });

package/serialization/resources/tts/types/PhonemeTimestamps.d.ts

@@ -0,0 +1,14 @@
+/**
+ * This file was auto-generated by Fern from our API Definition.
+ */
+import * as serializers from "../../../index";
+import * as Cartesia from "../../../../api/index";
+import * as core from "../../../../core";
+export declare const PhonemeTimestamps: core.serialization.ObjectSchema<serializers.PhonemeTimestamps.Raw, Cartesia.PhonemeTimestamps>;
+export declare namespace PhonemeTimestamps {
+    interface Raw {
+        phonemes: string[];
+        start: number[];
+        end: number[];
+    }
+}

package/serialization/resources/tts/types/PhonemeTimestamps.js

@@ -0,0 +1,35 @@
+"use strict";
+/**
+ * This file was auto-generated by Fern from our API Definition.
+ */
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || function (mod) {
+    if (mod && mod.__esModule) return mod;
+    var result = {};
+    if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
+    __setModuleDefault(result, mod);
+    return result;
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.PhonemeTimestamps = void 0;
+const core = __importStar(require("../../../../core"));
+exports.PhonemeTimestamps = core.serialization.object({
+    phonemes: core.serialization.list(core.serialization.string()),
+    start: core.serialization.list(core.serialization.number()),
+    end: core.serialization.list(core.serialization.number()),
+});

package/serialization/resources/tts/types/WebSocketPhonemeTimestampsResponse.d.ts

@@ -0,0 +1,14 @@
+/**
+ * This file was auto-generated by Fern from our API Definition.
+ */
+import * as serializers from "../../../index";
+import * as Cartesia from "../../../../api/index";
+import * as core from "../../../../core";
+import { PhonemeTimestamps } from "./PhonemeTimestamps";
+import { WebSocketBaseResponse } from "./WebSocketBaseResponse";
+export declare const WebSocketPhonemeTimestampsResponse: core.serialization.ObjectSchema<serializers.WebSocketPhonemeTimestampsResponse.Raw, Cartesia.WebSocketPhonemeTimestampsResponse>;
+export declare namespace WebSocketPhonemeTimestampsResponse {
+    interface Raw extends WebSocketBaseResponse.Raw {
+        phoneme_timestamps?: PhonemeTimestamps.Raw | null;
+    }
+}

package/serialization/resources/tts/types/WebSocketPhonemeTimestampsResponse.js

@@ -0,0 +1,37 @@
+"use strict";
+/**
+ * This file was auto-generated by Fern from our API Definition.
+ */
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || function (mod) {
+    if (mod && mod.__esModule) return mod;
+    var result = {};
+    if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
+    __setModuleDefault(result, mod);
+    return result;
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.WebSocketPhonemeTimestampsResponse = void 0;
+const core = __importStar(require("../../../../core"));
+const PhonemeTimestamps_1 = require("./PhonemeTimestamps");
+const WebSocketBaseResponse_1 = require("./WebSocketBaseResponse");
+exports.WebSocketPhonemeTimestampsResponse = core.serialization
+    .object({
+    phonemeTimestamps: core.serialization.property("phoneme_timestamps", PhonemeTimestamps_1.PhonemeTimestamps.optional()),
+})
+    .extend(WebSocketBaseResponse_1.WebSocketBaseResponse);

package/serialization/resources/tts/types/WebSocketResponse.d.ts

@@ -8,9 +8,10 @@ import { WebSocketChunkResponse } from "./WebSocketChunkResponse";
 import { WebSocketDoneResponse } from "./WebSocketDoneResponse";
 import { WebSocketTimestampsResponse } from "./WebSocketTimestampsResponse";
 import { WebSocketErrorResponse } from "./WebSocketErrorResponse";
+import { WebSocketPhonemeTimestampsResponse } from "./WebSocketPhonemeTimestampsResponse";
 export declare const WebSocketResponse: core.serialization.Schema<serializers.WebSocketResponse.Raw, Cartesia.WebSocketResponse>;
 export declare namespace WebSocketResponse {
-    type Raw = WebSocketResponse.Chunk | WebSocketResponse.Done | WebSocketResponse.Timestamps | WebSocketResponse.Error;
+    type Raw = WebSocketResponse.Chunk | WebSocketResponse.Done | WebSocketResponse.Timestamps | WebSocketResponse.Error | WebSocketResponse.PhonemeTimestamps;
     interface Chunk extends WebSocketChunkResponse.Raw {
         type: "chunk";
     }
@@ -23,4 +24,7 @@ export declare namespace WebSocketResponse {
     interface Error extends WebSocketErrorResponse.Raw {
         type: "error";
     }
+    interface PhonemeTimestamps extends WebSocketPhonemeTimestampsResponse.Raw {
+        type: "phoneme_timestamps";
+    }
 }

package/serialization/resources/tts/types/WebSocketResponse.js

@@ -32,12 +32,14 @@ const WebSocketChunkResponse_1 = require("./WebSocketChunkResponse");
 const WebSocketDoneResponse_1 = require("./WebSocketDoneResponse");
 const WebSocketTimestampsResponse_1 = require("./WebSocketTimestampsResponse");
 const WebSocketErrorResponse_1 = require("./WebSocketErrorResponse");
+const WebSocketPhonemeTimestampsResponse_1 = require("./WebSocketPhonemeTimestampsResponse");
 exports.WebSocketResponse = core.serialization
     .union("type", {
     chunk: WebSocketChunkResponse_1.WebSocketChunkResponse,
     done: WebSocketDoneResponse_1.WebSocketDoneResponse,
     timestamps: WebSocketTimestampsResponse_1.WebSocketTimestampsResponse,
     error: WebSocketErrorResponse_1.WebSocketErrorResponse,
+    phoneme_timestamps: WebSocketPhonemeTimestampsResponse_1.WebSocketPhonemeTimestampsResponse,
 })
     .transform({
     transform: (value) => value,

package/serialization/resources/tts/types/WebSocketTtsOutput.d.ts

@@ -5,11 +5,13 @@ import * as serializers from "../../../index";
 import * as Cartesia from "../../../../api/index";
 import * as core from "../../../../core";
 import { WordTimestamps } from "./WordTimestamps";
+import { PhonemeTimestamps } from "./PhonemeTimestamps";
 import { ContextId } from "./ContextId";
 export declare const WebSocketTtsOutput: core.serialization.ObjectSchema<serializers.WebSocketTtsOutput.Raw, Cartesia.WebSocketTtsOutput>;
 export declare namespace WebSocketTtsOutput {
     interface Raw {
         word_timestamps?: WordTimestamps.Raw | null;
+        phoneme_timestamps?: PhonemeTimestamps.Raw | null;
         audio?: unknown | null;
         context_id?: ContextId.Raw | null;
     }

package/serialization/resources/tts/types/WebSocketTtsOutput.js

@@ -29,9 +29,11 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.WebSocketTtsOutput = void 0;
 const core = __importStar(require("../../../../core"));
 const WordTimestamps_1 = require("./WordTimestamps");
+const PhonemeTimestamps_1 = require("./PhonemeTimestamps");
 const ContextId_1 = require("./ContextId");
 exports.WebSocketTtsOutput = core.serialization.object({
     wordTimestamps: core.serialization.property("word_timestamps", WordTimestamps_1.WordTimestamps.optional()),
+    phonemeTimestamps: core.serialization.property("phoneme_timestamps", PhonemeTimestamps_1.PhonemeTimestamps.optional()),
     audio: core.serialization.unknown().optional(),
     contextId: core.serialization.property("context_id", ContextId_1.ContextId.optional()),
 });

package/serialization/resources/tts/types/WebSocketTtsRequest.d.ts

@@ -16,6 +16,8 @@ export declare namespace WebSocketTtsRequest {
         duration?: number | null;
         language?: string | null;
         add_timestamps?: boolean | null;
+        add_phoneme_timestamps?: boolean | null;
+        continue?: boolean | null;
         context_id?: string | null;
     }
 }

package/serialization/resources/tts/types/WebSocketTtsRequest.js

@@ -38,5 +38,7 @@ exports.WebSocketTtsRequest = core.serialization.object({
     duration: core.serialization.number().optional(),
     language: core.serialization.string().optional(),
     addTimestamps: core.serialization.property("add_timestamps", core.serialization.boolean().optional()),
+    addPhonemeTimestamps: core.serialization.property("add_phoneme_timestamps", core.serialization.boolean().optional()),
+    continue: core.serialization.boolean().optional(),
     contextId: core.serialization.property("context_id", core.serialization.string().optional()),
 });

package/serialization/resources/tts/types/index.d.ts

@@ -4,9 +4,11 @@ export * from "./WebSocketResponse";
 export * from "./WebSocketErrorResponse";
 export * from "./WebSocketChunkResponse";
 export * from "./WebSocketTimestampsResponse";
+export * from "./WebSocketPhonemeTimestampsResponse";
 export * from "./WebSocketTtsOutput";
 export * from "./WebSocketStreamOptions";
 export * from "./WordTimestamps";
+export * from "./PhonemeTimestamps";
 export * from "./WebSocketDoneResponse";
 export * from "./CancelContextRequest";
 export * from "./GenerationRequest";

package/serialization/resources/tts/types/index.js

@@ -20,9 +20,11 @@ __exportStar(require("./WebSocketResponse"), exports);
 __exportStar(require("./WebSocketErrorResponse"), exports);
 __exportStar(require("./WebSocketChunkResponse"), exports);
 __exportStar(require("./WebSocketTimestampsResponse"), exports);
+__exportStar(require("./WebSocketPhonemeTimestampsResponse"), exports);
 __exportStar(require("./WebSocketTtsOutput"), exports);
 __exportStar(require("./WebSocketStreamOptions"), exports);
 __exportStar(require("./WordTimestamps"), exports);
+__exportStar(require("./PhonemeTimestamps"), exports);
 __exportStar(require("./WebSocketDoneResponse"), exports);
 __exportStar(require("./CancelContextRequest"), exports);
 __exportStar(require("./GenerationRequest"), exports);

package/serialization/resources/voices/types/CreateVoiceRequest.d.ts

@@ -6,7 +6,6 @@ import * as Cartesia from "../../../../api/index";
 import * as core from "../../../../core";
 import { Embedding } from "../../embedding/types/Embedding";
 import { SupportedLanguage } from "../../tts/types/SupportedLanguage";
-import { BaseVoiceId } from "./BaseVoiceId";
 export declare const CreateVoiceRequest: core.serialization.ObjectSchema<serializers.CreateVoiceRequest.Raw, Cartesia.CreateVoiceRequest>;
 export declare namespace CreateVoiceRequest {
     interface Raw {
@@ -14,6 +13,5 @@ export declare namespace CreateVoiceRequest {
         description: string;
         embedding: Embedding.Raw;
         language?: SupportedLanguage.Raw | null;
-        base_voice_id?: BaseVoiceId.Raw | null;
     }
 }

package/serialization/resources/voices/types/CreateVoiceRequest.js

@@ -30,11 +30,9 @@ exports.CreateVoiceRequest = void 0;
 const core = __importStar(require("../../../../core"));
 const Embedding_1 = require("../../embedding/types/Embedding");
 const SupportedLanguage_1 = require("../../tts/types/SupportedLanguage");
-const BaseVoiceId_1 = require("./BaseVoiceId");
 exports.CreateVoiceRequest = core.serialization.object({
     name: core.serialization.string(),
     description: core.serialization.string(),
     embedding: Embedding_1.Embedding,
     language: SupportedLanguage_1.SupportedLanguage.optional(),
-    baseVoiceId: core.serialization.property("base_voice_id", BaseVoiceId_1.BaseVoiceId.optional()),
 });

package/serialization/resources/voices/types/Voice.d.ts

@@ -7,7 +7,6 @@ import * as core from "../../../../core";
 import { VoiceId } from "./VoiceId";
 import { Embedding } from "../../embedding/types/Embedding";
 import { SupportedLanguage } from "../../tts/types/SupportedLanguage";
-import { BaseVoiceId } from "./BaseVoiceId";
 export declare const Voice: core.serialization.ObjectSchema<serializers.Voice.Raw, Cartesia.Voice>;
 export declare namespace Voice {
     interface Raw {
@@ -19,6 +18,5 @@ export declare namespace Voice {
         created_at: string;
         embedding: Embedding.Raw;
         language: SupportedLanguage.Raw;
-        base_voice_id?: BaseVoiceId.Raw | null;
     }
 }

package/serialization/resources/voices/types/Voice.js

@@ -31,7 +31,6 @@ const core = __importStar(require("../../../../core"));
 const VoiceId_1 = require("./VoiceId");
 const Embedding_1 = require("../../embedding/types/Embedding");
 const SupportedLanguage_1 = require("../../tts/types/SupportedLanguage");
-const BaseVoiceId_1 = require("./BaseVoiceId");
 exports.Voice = core.serialization.object({
     id: VoiceId_1.VoiceId,
     userId: core.serialization.property("user_id", core.serialization.string().optional()),
@@ -41,5 +40,4 @@ exports.Voice = core.serialization.object({
     createdAt: core.serialization.property("created_at", core.serialization.date()),
     embedding: Embedding_1.Embedding,
     language: SupportedLanguage_1.SupportedLanguage,
-    baseVoiceId: core.serialization.property("base_voice_id", BaseVoiceId_1.BaseVoiceId.optional()),
 });

package/version.d.ts

@@ -1 +1 @@
-export declare const SDK_VERSION = "2.1.2";
+export declare const SDK_VERSION = "2.1.4";

package/version.js

@@ -1,4 +1,4 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.SDK_VERSION = void 0;
-exports.SDK_VERSION = "2.1.2";
+exports.SDK_VERSION = "2.1.4";