@cartesia/cartesia-js 2.1.5 → 2.1.6

package/dist/api/resources/voices/client/Client.d.ts

@@ -40,24 +40,42 @@ export declare class Voices {
      */
     list(requestOptions?: Voices.RequestOptions): Promise<Cartesia.Voice[]>;
     /**
-     * @param {Cartesia.CreateVoiceRequest} request
+     * Clone a voice from an audio clip. This endpoint has two modes, stability and similarity.
+     *
+     * Similarity mode clones are more similar to the source clip, but may reproduce background noise. For these, use an audio clip about 5 seconds long.
+     *
+     * Stability mode clones are more stable, but may not sound as similar to the source clip. For these, use an audio clip 10-20 seconds long.
+     *
+     * @param {File | fs.ReadStream | Blob} clip
+     * @param {Cartesia.CloneVoiceRequest} request
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.create({
-     *         name: "string",
-     *         description: "string",
-     *         embedding: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
-     *         language: "en"
+     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
+     *         name: "A high-stability cloned voice",
+     *         description: "Copied from Cartesia docs",
+     *         mode: "stability",
+     *         language: "en",
+     *         enhance: true
+     *     })
+     *
+     * @example
+     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
+     *         name: "A high-similarity cloned voice",
+     *         description: "Copied from Cartesia docs",
+     *         mode: "similarity",
+     *         language: "en",
+     *         transcript: "A transcript of the words spoken in the audio clip.",
+     *         enhance: false
      *     })
      */
-    create(request: Cartesia.CreateVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.Voice>;
+    clone(clip: File | fs.ReadStream | Blob, request: Cartesia.CloneVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.VoiceMetadata>;
     /**
      * @param {Cartesia.VoiceId} id
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.delete("string")
+     *     await client.voices.delete("id")
      */
     delete(id: Cartesia.VoiceId, requestOptions?: Voices.RequestOptions): Promise<void>;
     /**
@@ -66,9 +84,9 @@ export declare class Voices {
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.update("string", {
-     *         name: "string",
-     *         description: "string"
+     *     await client.voices.update("id", {
+     *         name: "name",
+     *         description: "description"
      *     })
      */
     update(id: Cartesia.VoiceId, request: Cartesia.UpdateVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.Voice>;
@@ -77,7 +95,7 @@ export declare class Voices {
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.get("string")
+     *     await client.voices.get("id")
      */
     get(id: Cartesia.VoiceId, requestOptions?: Voices.RequestOptions): Promise<Cartesia.Voice>;
     /**
@@ -86,10 +104,10 @@ export declare class Voices {
      *
      * @example
      *     await client.voices.localize({
-     *         embedding: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
+     *         embedding: [1.1, 1.1],
      *         language: "en",
      *         originalSpeakerGender: "male",
-     *         dialect: "au"
+     *         dialect: undefined
      *     })
      */
     localize(request: Cartesia.LocalizeVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.EmbeddingResponse>;
@@ -100,43 +118,31 @@ export declare class Voices {
      * @example
      *     await client.voices.mix({
      *         voices: [{
-     *                 id: "string",
+     *                 id: "id",
+     *                 weight: 1.1
+     *             }, {
+     *                 id: "id",
      *                 weight: 1.1
      *             }]
      *     })
      */
     mix(request: Cartesia.MixVoicesRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.EmbeddingResponse>;
     /**
-     * Clone a voice from an audio clip. This endpoint has two modes, stability and similarity.
-     *
-     * Similarity mode clones are more similar to the source clip, but may reproduce background noise. For these, use an audio clip about 5 seconds long.
+     * Create voice from raw features. If you'd like to clone a voice from an audio file, please use Clone Voice instead.
      *
-     * Stability mode clones are more stable, but may not sound as similar to the source clip. For these, use an audio clip 10-20 seconds long.
-     *
-     * @param {File | fs.ReadStream | Blob} clip
-     * @param {Cartesia.CloneVoiceRequest} request
+     * @param {Cartesia.CreateVoiceRequest} request
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
-     *         name: "A high-stability cloned voice",
-     *         description: "Copied from Cartesia docs",
-     *         mode: "stability",
-     *         language: "en",
-     *         enhance: true
-     *     })
-     *
-     * @example
-     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
-     *         name: "A high-similarity cloned voice",
-     *         description: "Copied from Cartesia docs",
-     *         mode: "similarity",
+     *     await client.voices.create({
+     *         name: "My Custom Voice",
+     *         description: "A custom voice created through the API",
+     *         embedding: [],
      *         language: "en",
-     *         transcript: "A transcript of the words spoken in the audio clip.",
-     *         enhance: false
+     *         baseVoiceId: "123e4567-e89b-12d3-a456-426614174000"
      *     })
      */
-    clone(clip: File | fs.ReadStream | Blob, request: Cartesia.CloneVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.VoiceMetadata>;
+    create(request: Cartesia.CreateVoiceRequest, requestOptions?: Voices.RequestOptions): Promise<Cartesia.Voice>;
     protected _getCustomAuthorizationHeaders(): Promise<{
         "X-API-Key": string | undefined;
     }>;

package/dist/api/resources/voices/client/Client.js

@@ -70,7 +70,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/"),
                 method: "GET",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 timeoutMs: (requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.timeoutInSeconds) != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
@@ -108,33 +108,64 @@ class Voices {
         });
     }
     /**
-     * @param {Cartesia.CreateVoiceRequest} request
+     * Clone a voice from an audio clip. This endpoint has two modes, stability and similarity.
+     *
+     * Similarity mode clones are more similar to the source clip, but may reproduce background noise. For these, use an audio clip about 5 seconds long.
+     *
+     * Stability mode clones are more stable, but may not sound as similar to the source clip. For these, use an audio clip 10-20 seconds long.
+     *
+     * @param {File | fs.ReadStream | Blob} clip
+     * @param {Cartesia.CloneVoiceRequest} request
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.create({
-     *         name: "string",
-     *         description: "string",
-     *         embedding: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
-     *         language: "en"
+     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
+     *         name: "A high-stability cloned voice",
+     *         description: "Copied from Cartesia docs",
+     *         mode: "stability",
+     *         language: "en",
+     *         enhance: true
+     *     })
+     *
+     * @example
+     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
+     *         name: "A high-similarity cloned voice",
+     *         description: "Copied from Cartesia docs",
+     *         mode: "similarity",
+     *         language: "en",
+     *         transcript: "A transcript of the words spoken in the audio clip.",
+     *         enhance: false
      *     })
      */
-    create(request, requestOptions) {
+    clone(clip, request, requestOptions) {
         return __awaiter(this, void 0, void 0, function* () {
             var _a, _b, _c, _d, _e, _f;
+            const _request = yield core.newFormData();
+            yield _request.appendFile("clip", clip);
+            _request.append("name", request.name);
+            if (request.description != null) {
+                _request.append("description", request.description);
+            }
+            _request.append("language", serializers.SupportedLanguage.jsonOrThrow(request.language, { unrecognizedObjectKeys: "strip" }));
+            _request.append("mode", serializers.CloneMode.jsonOrThrow(request.mode, { unrecognizedObjectKeys: "strip" }));
+            _request.append("enhance", request.enhance.toString());
+            if (request.transcript != null) {
+                _request.append("transcript", request.transcript);
+            }
+            const _maybeEncodedRequest = yield _request.getRequest();
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
-                url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/"),
+                url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/clone"),
                 method: "POST",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
-                contentType: "application/json",
-                requestType: "json",
-                body: serializers.CreateVoiceRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
+                headers: Object.assign(Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), _maybeEncodedRequest.headers), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                requestType: "file",
+                duplex: _maybeEncodedRequest.duplex,
+                body: _maybeEncodedRequest.body,
                 timeoutMs: (requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.timeoutInSeconds) != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
                 maxRetries: requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.maxRetries,
                 abortSignal: requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.abortSignal,
             });
             if (_response.ok) {
-                return serializers.Voice.parseOrThrow(_response.body, {
+                return serializers.VoiceMetadata.parseOrThrow(_response.body, {
                     unrecognizedObjectKeys: "passthrough",
                     allowUnrecognizedUnionMembers: true,
                     allowUnrecognizedEnumValues: true,
@@ -155,7 +186,7 @@ class Voices {
                         body: _response.error.rawBody,
                     });
                 case "timeout":
-                    throw new errors.CartesiaTimeoutError("Timeout exceeded when calling POST /voices/.");
+                    throw new errors.CartesiaTimeoutError("Timeout exceeded when calling POST /voices/clone.");
                 case "unknown":
                     throw new errors.CartesiaError({
                         message: _response.error.errorMessage,
@@ -168,7 +199,7 @@ class Voices {
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.delete("string")
+     *     await client.voices.delete("id")
      */
     delete(id, requestOptions) {
         return __awaiter(this, void 0, void 0, function* () {
@@ -176,7 +207,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, `/voices/${encodeURIComponent(serializers.VoiceId.jsonOrThrow(id))}`),
                 method: "DELETE",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 timeoutMs: (requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.timeoutInSeconds) != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
@@ -213,9 +244,9 @@ class Voices {
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.update("string", {
-     *         name: "string",
-     *         description: "string"
+     *     await client.voices.update("id", {
+     *         name: "name",
+     *         description: "description"
      *     })
      */
     update(id, request, requestOptions) {
@@ -224,7 +255,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, `/voices/${encodeURIComponent(serializers.VoiceId.jsonOrThrow(id))}`),
                 method: "PATCH",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 body: serializers.UpdateVoiceRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
@@ -267,7 +298,7 @@ class Voices {
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.get("string")
+     *     await client.voices.get("id")
      */
     get(id, requestOptions) {
         return __awaiter(this, void 0, void 0, function* () {
@@ -275,7 +306,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, `/voices/${encodeURIComponent(serializers.VoiceId.jsonOrThrow(id))}`),
                 method: "GET",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 timeoutMs: (requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.timeoutInSeconds) != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
@@ -318,10 +349,10 @@ class Voices {
      *
      * @example
      *     await client.voices.localize({
-     *         embedding: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
+     *         embedding: [1.1, 1.1],
      *         language: "en",
      *         originalSpeakerGender: "male",
-     *         dialect: "au"
+     *         dialect: undefined
      *     })
      */
     localize(request, requestOptions) {
@@ -330,7 +361,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/localize"),
                 method: "POST",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 body: serializers.LocalizeVoiceRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
@@ -375,7 +406,10 @@ class Voices {
      * @example
      *     await client.voices.mix({
      *         voices: [{
-     *                 id: "string",
+     *                 id: "id",
+     *                 weight: 1.1
+     *             }, {
+     *                 id: "id",
      *                 weight: 1.1
      *             }]
      *     })
@@ -386,7 +420,7 @@ class Voices {
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
                 url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/mix"),
                 method: "POST",
-                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
                 contentType: "application/json",
                 requestType: "json",
                 body: serializers.MixVoicesRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
@@ -425,64 +459,36 @@ class Voices {
         });
     }
     /**
-     * Clone a voice from an audio clip. This endpoint has two modes, stability and similarity.
-     *
-     * Similarity mode clones are more similar to the source clip, but may reproduce background noise. For these, use an audio clip about 5 seconds long.
+     * Create voice from raw features. If you'd like to clone a voice from an audio file, please use Clone Voice instead.
      *
-     * Stability mode clones are more stable, but may not sound as similar to the source clip. For these, use an audio clip 10-20 seconds long.
-     *
-     * @param {File | fs.ReadStream | Blob} clip
-     * @param {Cartesia.CloneVoiceRequest} request
+     * @param {Cartesia.CreateVoiceRequest} request
      * @param {Voices.RequestOptions} requestOptions - Request-specific configuration.
      *
      * @example
-     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
-     *         name: "A high-stability cloned voice",
-     *         description: "Copied from Cartesia docs",
-     *         mode: "stability",
-     *         language: "en",
-     *         enhance: true
-     *     })
-     *
-     * @example
-     *     await client.voices.clone(fs.createReadStream("/path/to/your/file"), {
-     *         name: "A high-similarity cloned voice",
-     *         description: "Copied from Cartesia docs",
-     *         mode: "similarity",
+     *     await client.voices.create({
+     *         name: "My Custom Voice",
+     *         description: "A custom voice created through the API",
+     *         embedding: [],
      *         language: "en",
-     *         transcript: "A transcript of the words spoken in the audio clip.",
-     *         enhance: false
+     *         baseVoiceId: "123e4567-e89b-12d3-a456-426614174000"
      *     })
      */
-    clone(clip, request, requestOptions) {
+    create(request, requestOptions) {
         return __awaiter(this, void 0, void 0, function* () {
             var _a, _b, _c, _d, _e, _f;
-            const _request = yield core.newFormData();
-            yield _request.appendFile("clip", clip);
-            _request.append("name", request.name);
-            if (request.description != null) {
-                _request.append("description", request.description);
-            }
-            _request.append("language", serializers.SupportedLanguage.jsonOrThrow(request.language, { unrecognizedObjectKeys: "strip" }));
-            _request.append("mode", serializers.CloneMode.jsonOrThrow(request.mode, { unrecognizedObjectKeys: "strip" }));
-            _request.append("enhance", request.enhance.toString());
-            if (request.transcript != null) {
-                _request.append("transcript", request.transcript);
-            }
-            const _maybeEncodedRequest = yield _request.getRequest();
             const _response = yield ((_a = this._options.fetcher) !== null && _a !== void 0 ? _a : core.fetcher)({
-                url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/clone"),
+                url: (0, url_join_1.default)((_c = (_b = (yield core.Supplier.get(this._options.baseUrl))) !== null && _b !== void 0 ? _b : (yield core.Supplier.get(this._options.environment))) !== null && _c !== void 0 ? _c : environments.CartesiaEnvironment.Production, "/voices/"),
                 method: "POST",
-                headers: Object.assign(Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.5", "User-Agent": "@cartesia/cartesia-js/2.1.5", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), _maybeEncodedRequest.headers), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
-                requestType: "file",
-                duplex: _maybeEncodedRequest.duplex,
-                body: _maybeEncodedRequest.body,
+                headers: Object.assign(Object.assign({ "Cartesia-Version": (_f = (_d = requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.cartesiaVersion) !== null && _d !== void 0 ? _d : (_e = this._options) === null || _e === void 0 ? void 0 : _e.cartesiaVersion) !== null && _f !== void 0 ? _f : "2024-06-10", "X-Fern-Language": "JavaScript", "X-Fern-SDK-Name": "@cartesia/cartesia-js", "X-Fern-SDK-Version": "2.1.6", "User-Agent": "@cartesia/cartesia-js/2.1.6", "X-Fern-Runtime": core.RUNTIME.type, "X-Fern-Runtime-Version": core.RUNTIME.version }, (yield this._getCustomAuthorizationHeaders())), requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.headers),
+                contentType: "application/json",
+                requestType: "json",
+                body: serializers.CreateVoiceRequest.jsonOrThrow(request, { unrecognizedObjectKeys: "strip" }),
                 timeoutMs: (requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.timeoutInSeconds) != null ? requestOptions.timeoutInSeconds * 1000 : 60000,
                 maxRetries: requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.maxRetries,
                 abortSignal: requestOptions === null || requestOptions === void 0 ? void 0 : requestOptions.abortSignal,
             });
             if (_response.ok) {
-                return serializers.VoiceMetadata.parseOrThrow(_response.body, {
+                return serializers.Voice.parseOrThrow(_response.body, {
                     unrecognizedObjectKeys: "passthrough",
                     allowUnrecognizedUnionMembers: true,
                     allowUnrecognizedEnumValues: true,
@@ -503,7 +509,7 @@ class Voices {
                         body: _response.error.rawBody,
                     });
                 case "timeout":
-                    throw new errors.CartesiaTimeoutError("Timeout exceeded when calling POST /voices/clone.");
+                    throw new errors.CartesiaTimeoutError("Timeout exceeded when calling POST /voices/.");
                 case "unknown":
                     throw new errors.CartesiaError({
                         message: _response.error.errorMessage,

package/dist/api/resources/voices/types/BaseVoiceId.d.ts

@@ -3,6 +3,6 @@
  */
 import * as Cartesia from "../../../index";
 /**
- * The ID of the base voice associated with the voice, used for features like voice mixing.
+ * Pull in features from a base voice, used for features like voice mixing.
  */
 export type BaseVoiceId = Cartesia.VoiceId;

package/dist/api/resources/voices/types/CreateVoiceRequest.d.ts

@@ -9,4 +9,5 @@ export interface CreateVoiceRequest {
     description: string;
     embedding: Cartesia.Embedding;
     language?: Cartesia.SupportedLanguage;
+    baseVoiceId?: Cartesia.BaseVoiceId;
 }

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

@@ -6,6 +6,7 @@ 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 {
@@ -13,5 +14,6 @@ export declare namespace CreateVoiceRequest {
         description: string;
         embedding: Embedding.Raw;
         language?: SupportedLanguage.Raw | null;
+        base_voice_id?: BaseVoiceId.Raw | null;
     }
 }

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

@@ -40,9 +40,11 @@ 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/dist/version.d.ts

@@ -1 +1 @@
-export declare const SDK_VERSION = "2.1.5";
+export declare const SDK_VERSION = "2.1.6";

package/dist/version.js

@@ -1,4 +1,4 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.SDK_VERSION = void 0;
-exports.SDK_VERSION = "2.1.5";
+exports.SDK_VERSION = "2.1.6";

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@cartesia/cartesia-js",
-    "version": "2.1.5",
+    "version": "2.1.6",
     "private": false,
     "repository": "https://github.com/cartesia-ai/cartesia-js",
     "main": "./index.js",