speech-opencode 1.1.1 → 1.1.2

package/dist/index.d.ts

@@ -8,28 +8,9 @@ export interface VoicePluginOptions {
     silenceDuration?: number;
     /** Maximum recording duration in seconds as a safety timeout (default 300 = 5 minutes) */
     maxDuration?: number;
+    /** Enable wake word trigger file watching (default true) */
+    enableWakeWord?: boolean;
 }
-/**
- * OpenCode Voice Plugin
- *
- * Adds a 'voice' tool that records audio from the microphone and transcribes it
- * using OpenAI's Whisper API.
- *
- * @example
- * ```ts
- * // In opencode.json
- * {
- *   "plugin": ["opencode-voice"]
- * }
- * ```
- *
- * @example
- * ```ts
- * // With options in .opencode/plugin/voice.ts
- * import { VoicePlugin } from "opencode-voice"
- * export default VoicePlugin({ language: "en", defaultDuration: 10 })
- * ```
- */
 export declare const VoicePlugin: (options?: VoicePluginOptions) => Plugin;
 declare const _default: Plugin;
 export default _default;

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,MAAM,EAAQ,MAAM,qBAAqB,CAAA;AAwGvD,MAAM,WAAW,kBAAkB;IACjC,yDAAyD;IACzD,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,8FAA8F;IAC9F,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,+DAA+D;IAC/D,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,0FAA0F;IAC1F,WAAW,CAAC,EAAE,MAAM,CAAA;CACrB;AAED;;;;;;;;;;;;;;;;;;;;GAoBG;AACH,eAAO,MAAM,WAAW,GACrB,UAAS,kBAAuB,KAAG,MA4DnC,CAAA;;AAGH,wBAA4B"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,MAAM,EAAQ,MAAM,qBAAqB,CAAA;AA2GvD,MAAM,WAAW,kBAAkB;IACjC,yDAAyD;IACzD,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,8FAA8F;IAC9F,QAAQ,CAAC,EAAE,MAAM,CAAA;IACjB,+DAA+D;IAC/D,eAAe,CAAC,EAAE,MAAM,CAAA;IACxB,0FAA0F;IAC1F,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,4DAA4D;IAC5D,cAAc,CAAC,EAAE,OAAO,CAAA;CACzB;AAmID,eAAO,MAAM,WAAW,GACrB,UAAS,kBAAuB,KAAG,MAyDnC,CAAA;;AAGH,wBAA4B"}

package/dist/index.js

@@ -1,9 +1,11 @@
 import { tool } from "@opencode-ai/plugin";
 import OpenAI from "openai";
 import { spawn } from "child_process";
-import { unlinkSync, readFileSync } from "fs";
-import { tmpdir } from "os";
-import { join } from "path";
+import { unlinkSync, readFileSync, existsSync, watch, mkdirSync } from "fs";
+import { tmpdir, homedir } from "os";
+import { join, dirname } from "path";
+// Trigger file path for wake word integration
+const TRIGGER_FILE = join(homedir(), ".cache", "opencode", "voice_trigger");
 /**
  * Records audio from the microphone with automatic silence detection.
  * Recording stops after the specified silence duration.
@@ -99,11 +101,101 @@ async function transcribeAudio(audioFilePath, apiKey, language) {
  * export default VoicePlugin({ language: "en", defaultDuration: 10 })
  * ```
  */
+/**
+ * Records and transcribes audio, returning the transcription
+ */
+async function recordAndTranscribe(apiKey, maxDuration, silenceDuration, language) {
+    let audioFile = null;
+    try {
+        audioFile = await recordAudio(maxDuration, silenceDuration);
+        const transcription = await transcribeAudio(audioFile, apiKey, language);
+        if (!transcription || transcription.trim() === "") {
+            return "No speech detected. Please try again and speak clearly into your microphone.";
+        }
+        return transcription;
+    }
+    finally {
+        if (audioFile) {
+            try {
+                unlinkSync(audioFile);
+            }
+            catch {
+                // Ignore cleanup errors
+            }
+        }
+    }
+}
+/**
+ * Clears the wake word trigger file
+ */
+function clearTriggerFile() {
+    try {
+        if (existsSync(TRIGGER_FILE)) {
+            unlinkSync(TRIGGER_FILE);
+        }
+    }
+    catch {
+        // Ignore errors
+    }
+}
+/**
+ * Sets up wake word trigger file watching
+ * When the trigger file is written, it records audio, transcribes it, and appends to the TUI prompt
+ */
+function setupWakeWordWatcher(apiKey, maxDuration, silenceDuration, language, client // OpenCode SDK client
+) {
+    // Ensure the directory exists
+    const triggerDir = dirname(TRIGGER_FILE);
+    mkdirSync(triggerDir, { recursive: true });
+    // Clear any existing trigger
+    clearTriggerFile();
+    console.log("[Voice Plugin] Wake word watcher enabled");
+    console.log(`[Voice Plugin] Watching: ${TRIGGER_FILE}`);
+    console.log("[Voice Plugin] Run 'python wakeword/listener.py' to enable 'Hey Jarvis' wake word");
+    // Watch the directory for the trigger file
+    let isRecording = false;
+    watch(triggerDir, async (eventType, filename) => {
+        if (filename !== "voice_trigger" || isRecording)
+            return;
+        if (!existsSync(TRIGGER_FILE))
+            return;
+        isRecording = true;
+        console.log("[Voice Plugin] Wake word triggered! Recording...");
+        try {
+            // Clear the trigger file immediately
+            clearTriggerFile();
+            // Record and transcribe
+            const transcription = await recordAndTranscribe(apiKey, maxDuration, silenceDuration, language);
+            if (transcription && !transcription.startsWith("No speech detected")) {
+                console.log(`[Voice Plugin] Transcribed: "${transcription}"`);
+                // Append transcription to the TUI prompt
+                try {
+                    await client.tui.appendPrompt({ body: { text: transcription } });
+                    // Auto-submit the prompt
+                    await client.tui.submitPrompt();
+                }
+                catch (err) {
+                    console.error("[Voice Plugin] Failed to send to TUI:", err);
+                }
+            }
+        }
+        catch (error) {
+            console.error("[Voice Plugin] Error:", error);
+        }
+        finally {
+            isRecording = false;
+        }
+    });
+}
 export const VoicePlugin = (options = {}) => async (ctx) => {
-    const { apiKey = process.env.OPENAI_API_KEY, language, silenceDuration = 7, maxDuration = 300, } = options;
+    const { apiKey = process.env.OPENAI_API_KEY, language, silenceDuration = 7, maxDuration = 300, enableWakeWord = true, } = options;
     if (!apiKey) {
         console.warn("[Voice Plugin] Warning: OPENAI_API_KEY not set. Voice transcription will fail.");
     }
+    // Set up wake word watcher if enabled
+    if (enableWakeWord && apiKey && ctx.client) {
+        setupWakeWordWatcher(apiKey, maxDuration, silenceDuration, language, ctx.client);
+    }
     return {
         tool: {
             voice: tool({
@@ -115,29 +207,14 @@ export const VoicePlugin = (options = {}) => async (ctx) => {
                     if (!apiKey) {
                         return "Error: OPENAI_API_KEY environment variable is not set. Please set it to use voice transcription.";
                     }
-                    let audioFile = null;
                     try {
-                        audioFile = await recordAudio(maxDuration, silenceDuration);
-                        const transcription = await transcribeAudio(audioFile, apiKey, language);
-                        if (!transcription || transcription.trim() === "") {
-                            return "No speech detected. Please try again and speak clearly into your microphone.";
-                        }
+                        const transcription = await recordAndTranscribe(apiKey, maxDuration, silenceDuration, language);
                         return `Transcribed speech: "${transcription}"`;
                     }
                     catch (error) {
                         const errorMessage = error instanceof Error ? error.message : String(error);
                         return `Voice recording/transcription failed: ${errorMessage}`;
                     }
-                    finally {
-                        if (audioFile) {
-                            try {
-                                unlinkSync(audioFile);
-                            }
-                            catch {
-                                // Ignore cleanup errors
-                            }
-                        }
-                    }
                 },
             }),
         },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "speech-opencode",
-  "version": "1.1.1",
+  "version": "1.1.2",
   "description": "Voice input plugin for OpenCode using OpenAI Whisper",
   "keywords": [
     "opencode",