support inworld tts (#537)

* support inworld tts

* wip

src/api/constants.ts

@@ -3,6 +3,7 @@ import type {
   Currency,
   ElevenLabsOptions,
   GoogleCustomVoice,
+  InworldOptions,
   LimitField,
   LimitUnitOption,
   PasswordSettings,
@@ -277,6 +278,14 @@ export const DEFAULT_RIMELABS_OPTIONS: Partial<RimelabsOptions> = {
   reduceLatency: true,
 };
 
+export const DEFAULT_INWORLD_OPTIONS: Partial<InworldOptions> = {
+  audioConfig: {
+    pitch: 0.0,
+    speakingRate: 1.0,
+  },
+  temperature: 0.8,
+};
+
 // PlayHT options
 export const DEFAULT_PLAYHT_OPTIONS: Partial<PlayHTOptions> = {
   quality: "medium",

src/api/types.ts

@@ -781,6 +781,16 @@ export interface RimelabsOptions {
   reduceLatency: boolean;
 }
 
+export interface InworldOptions {
+  audioConfig: {
+    bitRate?: number;
+    sampleRateHertz?: number;
+    pitch?: number;
+    speakingRate?: number;
+  };
+  temperature?: number;
+}
+
 export type CartesiaEmotions =
   | "anger:lowest"
   | "anger:low"

src/containers/internal/views/applications/speech-selection.tsx

@@ -35,6 +35,7 @@ import {
   VENDOR_VOXIST,
   VENDOR_RIMELABS,
   VENDOR_OPENAI,
+  VENDOR_INWORLD,
 } from "src/vendor";
 import {
   LabelOptions,
@@ -312,6 +313,15 @@ export const SpeechProviderSelection = ({
             updateTtsVoice(newLang!.value, newLang!.voices[0].value);
             return;
           }
+          if (synthVendor === VENDOR_INWORLD) {
+            let newLang = json.tts.find((lang) => lang.value === "en");
+            // If the new language doesn't map then default to the first one
+            if (!newLang) {
+              newLang = json.tts[0];
+            }
+            updateTtsVoice(newLang!.value, newLang!.voices[0].value);
+            return;
+          }
           /** Google and AWS have different language lists */
           /** If the new language doesn't map then default to "en-US" */
           let newLang = json.tts.find((lang) => lang.value === synthLang);

src/containers/internal/views/speech-services/form.tsx

@@ -52,6 +52,7 @@ import {
   VENDOR_CARTESIA,
   VENDOR_VOXIST,
   VENDOR_OPENAI,
+  VENDOR_INWORLD,
 } from "src/vendor";
 import { MSG_REQUIRED_FIELDS } from "src/constants";
 import {
@@ -83,6 +84,7 @@ import {
   DEFAULT_CARTESIA_OPTIONS,
   DEFAULT_ELEVENLABS_OPTIONS,
   DEFAULT_GOOGLE_CUSTOM_VOICE,
+  DEFAULT_INWORLD_OPTIONS,
   DEFAULT_PLAYHT_OPTIONS,
   DEFAULT_RIMELABS_OPTIONS,
   DEFAULT_VERBIO_MODEL,
@@ -233,6 +235,8 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
           return DEFAULT_PLAYHT_OPTIONS;
         case VENDOR_RIMELABS:
           return DEFAULT_RIMELABS_OPTIONS;
+        case VENDOR_INWORLD:
+          return DEFAULT_INWORLD_OPTIONS;
         case VENDOR_CARTESIA:
           return DEFAULT_CARTESIA_OPTIONS;
       }
@@ -249,6 +253,8 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
           return "https://docs.play.ht/reference/api-generate-tts-audio-stream";
         case VENDOR_RIMELABS:
           return "https://rimelabs.mintlify.app/api-reference/endpoint/streaming-mp3#variable-parameters";
+        case VENDOR_INWORLD:
+          return "https://docs.inworld.ai/api-reference/ttsAPI/texttospeech/synthesize-speech-stream";
         case VENDOR_CARTESIA:
           return "https://docs.cartesia.ai/api-reference/tts/bytes";
       }
@@ -432,11 +438,13 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
         }),
         ...((vendor === VENDOR_ELEVENLABS ||
           vendor === VENDOR_WHISPER ||
+          vendor === VENDOR_INWORLD ||
           vendor === VENDOR_RIMELABS) && {
           model_id: ttsModelId || null,
         }),
         ...((vendor === VENDOR_ELEVENLABS ||
           vendor === VENDOR_PLAYHT ||
+          vendor === VENDOR_INWORLD ||
           vendor === VENDOR_RIMELABS) && {
           options: options || null,
         }),
@@ -509,6 +517,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
               vendor === VENDOR_ELEVENLABS ||
               vendor === VENDOR_PLAYHT ||
               vendor === VENDOR_RIMELABS ||
+              vendor === VENDOR_INWORLD ||
               vendor === VENDOR_WHISPER ||
               vendor === VENDOR_CARTESIA ||
               vendor === VENDOR_OPENAI
@@ -578,6 +587,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
       vendor === VENDOR_WHISPER ||
       vendor === VENDOR_PLAYHT ||
       vendor === VENDOR_RIMELABS ||
+      vendor === VENDOR_INWORLD ||
       vendor === VENDOR_CARTESIA ||
       vendor === VENDOR_OPENAI ||
       vendor === VENDOR_DEEPGRAM
@@ -976,6 +986,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
               vendor !== VENDOR_WHISPER &&
               vendor !== VENDOR_PLAYHT &&
               vendor !== VENDOR_RIMELABS &&
+              vendor !== VENDOR_INWORLD &&
               vendor !== VENDOR_ELEVENLABS && (
                 <label htmlFor="use_for_stt" className="chk">
                   <input
@@ -1705,6 +1716,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
           vendor == VENDOR_ELEVENLABS ||
           vendor === VENDOR_WHISPER ||
           vendor === VENDOR_RIMELABS ||
+          vendor === VENDOR_INWORLD ||
           vendor === VENDOR_SONIOX ||
           vendor === VENDOR_CARTESIA ||
           vendor === VENDOR_OPENAI ||
@@ -1724,10 +1736,11 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
             />
           </fieldset>
         )}
-        {(vendor == VENDOR_ELEVENLABS ||
+        {(vendor === VENDOR_ELEVENLABS ||
-          vendor == VENDOR_WHISPER ||
+          vendor === VENDOR_WHISPER ||
           vendor === VENDOR_PLAYHT ||
-          vendor == VENDOR_RIMELABS ||
+          vendor === VENDOR_RIMELABS ||
+          vendor === VENDOR_INWORLD ||
           (ttsCheck && vendor === VENDOR_CARTESIA)) &&
           ttsModels.length > 0 && (
             <fieldset>
@@ -1767,7 +1780,8 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
         {(vendor === VENDOR_ELEVENLABS ||
           vendor === VENDOR_PLAYHT ||
           vendor === VENDOR_CARTESIA ||
-          vendor === VENDOR_RIMELABS) && (
+          vendor === VENDOR_RIMELABS ||
+          vendor === VENDOR_INWORLD) && (
           <fieldset>
             <Checkzone
               hidden

src/vendor/index.tsx

@@ -24,6 +24,7 @@ export const VENDOR_VOXIST = "voxist";
 export const VENDOR_WHISPER = "whisper";
 export const VENDOR_PLAYHT = "playht";
 export const VENDOR_RIMELABS = "rimelabs";
+export const VENDOR_INWORLD = "inworld";
 export const VENDOR_VERBIO = "verbio";
 export const VENDOR_CARTESIA = "cartesia";
 export const VENDOR_OPENAI = "openai";
@@ -101,6 +102,10 @@ export const vendors: VendorOptions[] = [
     name: "RimeLabs",
     value: VENDOR_RIMELABS,
   },
+  {
+    name: "Inworld",
+    value: VENDOR_INWORLD,
+  },
   {
     name: "Verbio",
     value: VENDOR_VERBIO,

src/vendor/types.ts

@@ -17,6 +17,7 @@ export type Vendor =
   | "whisper"
   | "playht"
   | "rimelabs"
+  | "inworld"
   | "verbio"
   | "openai"
   | "Cartesia";
@@ -112,6 +113,7 @@ export interface SynthesisVendors {
   playht: VoiceLanguage[];
   cartesia: VoiceLanguage[];
   rimelabs: VoiceLanguage[];
+  inworld: VoiceLanguage[];
 }
 
 export interface MSRawSpeech {