phonic 0.6.1 → 0.8.0

package/README.md

@@ -17,7 +17,7 @@ npm i phonic
 
 ## Setup
 
-Grab an API key from [Phonic settings](https://phonic.co/settings) and pass it to the Phonic constructor.
+Grab an API key from the [Phonic API Keys](https://phonic.co/api-keys) section and pass it to the Phonic constructor.
 
 ```ts
 import { Phonic } from "phonic";
@@ -38,7 +38,7 @@ if (error === null) {
 ```
 
 
-### Get voice by id
+### Get voice by ID
 
 ```ts
 const { data, error } = await phonic.voices.get("meredith");
@@ -50,23 +50,10 @@ if (error === null) {
 
 ### Speech-to-speech via WebSocket
 
-Open a WebSocket connection:
+To start a conversation, open a WebSocket connection:
 
 ```ts
-const { data, error } = await phonic.sts.websocket();
-
-if (error !== null) {
-  throw new Error(error.message);
-}
-
-// Here we know that the WebSocket connection is open.
-const { phonicWebSocket } = data;
-```
-
-Send config params for the conversation:
-
-```ts
-phonicWebSocket.config({
+const { data, error } = await phonic.sts.websocket({
   input_format: "mulaw_8000",
 
   // Optional fields
@@ -75,6 +62,12 @@ phonicWebSocket.config({
   voice_id: "meredith",
   output_format: "mulaw_8000"
 });
+
+if (error !== null) {
+  throw new Error(`Failed to start conversation: ${error.message}`);
+}
+
+const { phonicWebSocket } = data;
 ```
 
 Stream input (user) audio chunks:
@@ -132,16 +125,6 @@ phonicWebSocket.onError((event) => {
 });
 ```
 
-## Publish a new version on npm
-
-1. `bunx changeset`
-2. `git add .`
-3. `git commit -m "Add changeset"`
-4. `git push`
-
-This should trigger the `publish` github workflow that will create a Pull Request named "Version Packages". 
-Once this Pull Request is merged, the new version will be published on npm.
-
 ## License
 
 MIT

package/dist/index.d.mts

@@ -26,6 +26,8 @@ type PhonicSTSConfig = {
     output_format?: "pcm_44100" | "mulaw_8000";
 };
 type PhonicSTSWebSocketResponseMessage = {
+    type: "ready_to_start_conversation";
+} | {
     type: "input_text";
     text: string;
 } | {
@@ -62,7 +64,6 @@ declare class PhonicSTSWebSocket {
     onMessage(callback: OnMessageCallback): void;
     onClose(callback: OnCloseCallback): void;
     onError(callback: OnErrorCallback): void;
-    config(message: PhonicSTSConfig): void;
     audioChunk(message: {
         audio: string;
     }): void;
@@ -72,7 +73,8 @@ declare class PhonicSTSWebSocket {
 declare class SpeechToSpeech {
     private readonly phonic;
     constructor(phonic: Phonic);
-    websocket(): DataOrError<{
+    private connectToPhonicAPI;
+    websocket(config: PhonicSTSConfig): DataOrError<{
         phonicWebSocket: PhonicSTSWebSocket;
     }>;
 }

package/dist/index.d.ts

@@ -26,6 +26,8 @@ type PhonicSTSConfig = {
     output_format?: "pcm_44100" | "mulaw_8000";
 };
 type PhonicSTSWebSocketResponseMessage = {
+    type: "ready_to_start_conversation";
+} | {
     type: "input_text";
     text: string;
 } | {
@@ -62,7 +64,6 @@ declare class PhonicSTSWebSocket {
     onMessage(callback: OnMessageCallback): void;
     onClose(callback: OnCloseCallback): void;
     onError(callback: OnErrorCallback): void;
-    config(message: PhonicSTSConfig): void;
     audioChunk(message: {
         audio: string;
     }): void;
@@ -72,7 +73,8 @@ declare class PhonicSTSWebSocket {
 declare class SpeechToSpeech {
     private readonly phonic;
     constructor(phonic: Phonic);
-    websocket(): DataOrError<{
+    private connectToPhonicAPI;
+    websocket(config: PhonicSTSConfig): DataOrError<{
         phonicWebSocket: PhonicSTSWebSocket;
     }>;
 }

package/dist/index.js

@@ -35,7 +35,7 @@ __export(index_exports, {
 module.exports = __toCommonJS(index_exports);
 
 // package.json
-var version = "0.6.1";
+var version = "0.8.0";
 
 // src/sts/index.ts
 var import_ws = __toESM(require("ws"));
@@ -71,7 +71,6 @@ var PhonicSTSWebSocket = class {
     this.onMessage = this.onMessage.bind(this);
     this.onClose = this.onClose.bind(this);
     this.onError = this.onError.bind(this);
-    this.config = this.config.bind(this);
     this.audioChunk = this.audioChunk.bind(this);
     this.close = this.close.bind(this);
   }
@@ -87,14 +86,6 @@ var PhonicSTSWebSocket = class {
   onError(callback) {
     this.onErrorCallback = callback;
   }
-  config(message) {
-    this.ws.send(
-      JSON.stringify({
-        type: "config",
-        ...message
-      })
-    );
-  }
   audioChunk(message) {
     this.ws.send(
       JSON.stringify({
@@ -109,21 +100,43 @@ var PhonicSTSWebSocket = class {
 };
 
 // src/sts/index.ts
+var phonicApiCloseCodes = {
+  insuffucientCapacityAvailable: 4004
+};
 var SpeechToSpeech = class {
   constructor(phonic) {
     this.phonic = phonic;
   }
-  async websocket() {
+  async connectToPhonicAPI(phonicApiWsUrl, config) {
     return new Promise((resolve) => {
-      const wsBaseUrl = this.phonic.baseUrl.replace(/^http/, "ws");
-      const ws = new import_ws.default(`${wsBaseUrl}/v1/sts/ws`, {
+      const ws = new import_ws.default(phonicApiWsUrl, {
         headers: {
           Authorization: `Bearer ${this.phonic.apiKey}`
         }
       });
       ws.onopen = () => {
-        const phonicWebSocket = new PhonicSTSWebSocket(ws);
-        resolve({ data: { phonicWebSocket }, error: null });
+        ws.send(
+          JSON.stringify({
+            type: "config",
+            ...config
+          })
+        );
+      };
+      ws.onmessage = (event) => {
+        if (typeof event.data !== "string") {
+          throw new Error("Received non-string message");
+        }
+        const dataObj = JSON.parse(
+          event.data
+        );
+        if (dataObj.type === "ready_to_start_conversation") {
+          resolve({
+            data: {
+              phonicWebSocket: new PhonicSTSWebSocket(ws)
+            },
+            error: null
+          });
+        }
       };
       ws.onerror = (error) => {
         resolve({
@@ -133,8 +146,48 @@ var SpeechToSpeech = class {
           }
         });
       };
+      ws.onclose = (event) => {
+        if (event.code === phonicApiCloseCodes.insuffucientCapacityAvailable) {
+          resolve({
+            data: null,
+            error: {
+              message: event.reason,
+              code: "insuffucient_capacity_available"
+            }
+          });
+        }
+      };
     });
   }
+  async websocket(config) {
+    const wsBaseUrl = this.phonic.baseUrl.replace(/^http/, "ws");
+    const phonicApiWsUrl = `${wsBaseUrl}/v1/sts/ws`;
+    let retryNumber = 0;
+    const maxRetries = 14;
+    const retryDelay = 15e3;
+    while (true) {
+      const connectResult = await this.connectToPhonicAPI(
+        phonicApiWsUrl,
+        config
+      );
+      if (connectResult.data !== null) {
+        return connectResult;
+      }
+      if (connectResult.error.code === "insuffucient_capacity_available") {
+        if (retryNumber >= maxRetries) {
+          return connectResult;
+        }
+        console.info(
+          `${connectResult.error.message}, will retry in ${retryDelay / 1e3}sec`
+        );
+        await new Promise((resolve) => setTimeout(resolve, retryDelay));
+        retryNumber += 1;
+        console.info(`Retrying... ${retryNumber}/${maxRetries}`);
+        continue;
+      }
+      return connectResult;
+    }
+  }
 };
 
 // src/voices/index.ts

package/dist/index.mjs

@@ -1,5 +1,5 @@
 // package.json
-var version = "0.6.1";
+var version = "0.8.0";
 
 // src/sts/index.ts
 import WebSocket from "ws";
@@ -35,7 +35,6 @@ var PhonicSTSWebSocket = class {
     this.onMessage = this.onMessage.bind(this);
     this.onClose = this.onClose.bind(this);
     this.onError = this.onError.bind(this);
-    this.config = this.config.bind(this);
     this.audioChunk = this.audioChunk.bind(this);
     this.close = this.close.bind(this);
   }
@@ -51,14 +50,6 @@ var PhonicSTSWebSocket = class {
   onError(callback) {
     this.onErrorCallback = callback;
   }
-  config(message) {
-    this.ws.send(
-      JSON.stringify({
-        type: "config",
-        ...message
-      })
-    );
-  }
   audioChunk(message) {
     this.ws.send(
       JSON.stringify({
@@ -73,21 +64,43 @@ var PhonicSTSWebSocket = class {
 };
 
 // src/sts/index.ts
+var phonicApiCloseCodes = {
+  insuffucientCapacityAvailable: 4004
+};
 var SpeechToSpeech = class {
   constructor(phonic) {
     this.phonic = phonic;
   }
-  async websocket() {
+  async connectToPhonicAPI(phonicApiWsUrl, config) {
     return new Promise((resolve) => {
-      const wsBaseUrl = this.phonic.baseUrl.replace(/^http/, "ws");
-      const ws = new WebSocket(`${wsBaseUrl}/v1/sts/ws`, {
+      const ws = new WebSocket(phonicApiWsUrl, {
         headers: {
           Authorization: `Bearer ${this.phonic.apiKey}`
         }
       });
       ws.onopen = () => {
-        const phonicWebSocket = new PhonicSTSWebSocket(ws);
-        resolve({ data: { phonicWebSocket }, error: null });
+        ws.send(
+          JSON.stringify({
+            type: "config",
+            ...config
+          })
+        );
+      };
+      ws.onmessage = (event) => {
+        if (typeof event.data !== "string") {
+          throw new Error("Received non-string message");
+        }
+        const dataObj = JSON.parse(
+          event.data
+        );
+        if (dataObj.type === "ready_to_start_conversation") {
+          resolve({
+            data: {
+              phonicWebSocket: new PhonicSTSWebSocket(ws)
+            },
+            error: null
+          });
+        }
       };
       ws.onerror = (error) => {
         resolve({
@@ -97,8 +110,48 @@ var SpeechToSpeech = class {
           }
         });
       };
+      ws.onclose = (event) => {
+        if (event.code === phonicApiCloseCodes.insuffucientCapacityAvailable) {
+          resolve({
+            data: null,
+            error: {
+              message: event.reason,
+              code: "insuffucient_capacity_available"
+            }
+          });
+        }
+      };
     });
   }
+  async websocket(config) {
+    const wsBaseUrl = this.phonic.baseUrl.replace(/^http/, "ws");
+    const phonicApiWsUrl = `${wsBaseUrl}/v1/sts/ws`;
+    let retryNumber = 0;
+    const maxRetries = 14;
+    const retryDelay = 15e3;
+    while (true) {
+      const connectResult = await this.connectToPhonicAPI(
+        phonicApiWsUrl,
+        config
+      );
+      if (connectResult.data !== null) {
+        return connectResult;
+      }
+      if (connectResult.error.code === "insuffucient_capacity_available") {
+        if (retryNumber >= maxRetries) {
+          return connectResult;
+        }
+        console.info(
+          `${connectResult.error.message}, will retry in ${retryDelay / 1e3}sec`
+        );
+        await new Promise((resolve) => setTimeout(resolve, retryDelay));
+        retryNumber += 1;
+        console.info(`Retrying... ${retryNumber}/${maxRetries}`);
+        continue;
+      }
+      return connectResult;
+    }
+  }
 };
 
 // src/voices/index.ts

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "phonic",
-  "version": "0.6.1",
+  "version": "0.8.0",
   "description": "Phonic Node.js SDK",
   "scripts": {
     "build": "tsup",
@@ -39,8 +39,8 @@
     "@biomejs/biome": "1.9.4",
     "@changesets/changelog-github": "0.5.1",
     "@changesets/cli": "2.28.1",
-    "@types/bun": "1.2.3",
-    "tsup": "8.3.6",
+    "@types/bun": "1.2.4",
+    "tsup": "8.4.0",
     "typescript": "5.7.3",
     "zod": "3.24.2"
   },