pi-voice 0.2.0 → 0.3.3

package/README.md

@@ -52,14 +52,14 @@ You can configure pi-voice in `.pi/pi-voice.json`:
 | Key | Description |
 | --- | --- |
 | `key` | Push-to-talk shortcut. Combine modifiers (`ctrl`, `shift`, `alt`/`opt`, `meta`/`cmd`) and a main key with `+`. Examples: `"ctrl+t"`, `"alt+space"`, `"ctrl+shift+r"`. Default: `"meta+shift+i"`. |
-| `provider` | Speech provider for STT & TTS. `"local"`, `"gemini"` (Vertex AI), or `"openai"`. Default: `"local"`. |
+| `provider` | Speech provider for STT & TTS. `"local"`, `"gemini"` (Vertex AI or Gemini API), or `"openai"`. Default: `"local"`. |
 
 ### Environment variables
 
 | Provider | Required variables |
 | --- | --- |
 | `local` | None (model is auto-downloaded on first launch). Optional: `WHISPER_MODEL_PATH` (custom model path), `WHISPER_MODEL` (model name, default `medium-q5_0`), `SAY_VOICE` (macOS `say` voice name, e.g. `"Kyoko"`). |
-| `gemini` | `GOOGLE_CLOUD_PROJECT`, `GOOGLE_CLOUD_LOCATION` (optional, default `us-central1`) |
+| `gemini` | **Vertex AI:** `GOOGLE_CLOUD_PROJECT`, `GOOGLE_CLOUD_LOCATION` (optional, default `us-central1`). **Gemini API:** `GEMINI_API_KEY` or `GOOGLE_API_KEY`. If `GOOGLE_CLOUD_PROJECT` is set, Vertex AI is used; set `GOOGLE_GENAI_USE_VERTEXAI=false` to force API key mode. |
 | `openai` | `OPENAI_API_KEY` |
 
 #### Logging

package/out/cli/cli.js

@@ -1952,7 +1952,7 @@ var require_indexes = __commonJS((exports, module) => {
 
 // node_modules/thread-stream/index.js
 var require_thread_stream = __commonJS((exports, module) => {
-  var __dirname = "/Users/yukukotani/ghq/github.com/yukukotani/pi-voice/node_modules/thread-stream";
+  var __dirname = "/home/runner/work/pi-voice/pi-voice/node_modules/thread-stream";
   var { version } = require_package();
   var { EventEmitter } = __require("events");
   var { Worker } = __require("worker_threads");
@@ -2377,7 +2377,7 @@ var require_thread_stream = __commonJS((exports, module) => {
 
 // node_modules/pino/lib/transport.js
 var require_transport = __commonJS((exports, module) => {
-  var __dirname = "/Users/yukukotani/ghq/github.com/yukukotani/pi-voice/node_modules/pino/lib";
+  var __dirname = "/home/runner/work/pi-voice/pi-voice/node_modules/pino/lib";
   var { createRequire: createRequire2 } = __require("module");
   var { existsSync: existsSync2 } = __require("node:fs");
   var getCallers = require_caller();
@@ -4492,7 +4492,7 @@ var require_node_gyp_build2 = __commonJS((exports, module) => {
 
 // node_modules/uiohook-napi/dist/index.js
 var require_dist = __commonJS((exports) => {
-  var __dirname = "/Users/yukukotani/ghq/github.com/yukukotani/pi-voice/node_modules/uiohook-napi/dist";
+  var __dirname = "/home/runner/work/pi-voice/pi-voice/node_modules/uiohook-napi/dist";
   Object.defineProperty(exports, "__esModule", { value: true });
   exports.uIOhook = exports.UiohookKey = exports.WheelDirection = exports.EventType = undefined;
   var events_1 = __require("events");

package/out/main/index.js

@@ -6,9 +6,9 @@ import { homedir } from "node:os";
 import pino from "pino";
 import { readFileSync, existsSync, mkdirSync, createWriteStream, unlinkSync, writeFileSync } from "node:fs";
 import { z } from "zod";
-import { GoogleGenAI } from "@google/genai";
 import OpenAI, { toFile } from "openai";
 import { WhisperFullParams, WhisperSamplingStrategy, Whisper } from "@napi-rs/whisper";
+import { GoogleGenAI } from "@google/genai";
 import { Readable } from "node:stream";
 import { finished } from "node:stream/promises";
 import { spawn } from "node:child_process";
@@ -302,6 +302,26 @@ function loadConfig(cwd) {
   logger.info({ key: display, provider: parsed.provider, configPath }, "Loaded config");
   return { key: binding, keyDisplay: display, provider: parsed.provider };
 }
+let geminiClient = null;
+function getGeminiClient() {
+  if (geminiClient) return geminiClient;
+  const forceVertexOff = process.env.GOOGLE_GENAI_USE_VERTEXAI === "false";
+  const project = process.env.GOOGLE_CLOUD_PROJECT;
+  const location = process.env.GOOGLE_CLOUD_LOCATION ?? "us-central1";
+  const apiKey = process.env.GEMINI_API_KEY ?? process.env.GOOGLE_API_KEY;
+  if (project && !forceVertexOff) {
+    logger.info({ project, location }, "Initializing Gemini client (Vertex AI)");
+    geminiClient = new GoogleGenAI({ vertexai: true, project, location });
+  } else if (apiKey) {
+    logger.info("Initializing Gemini client (API key)");
+    geminiClient = new GoogleGenAI({ apiKey });
+  } else {
+    throw new Error(
+      "Gemini provider requires either GOOGLE_CLOUD_PROJECT (for Vertex AI) or GEMINI_API_KEY / GOOGLE_API_KEY (for Gemini API)."
+    );
+  }
+  return geminiClient;
+}
 const DEFAULT_MODEL = "medium-q5_0";
 const HF_BASE_URL = "https://huggingface.co/ggerganov/whisper.cpp/resolve/main";
 function modelCacheDir() {
@@ -393,17 +413,6 @@ async function resolveModelPath() {
   await downloadModel(model, destPath);
   return destPath;
 }
-let geminiClient$1 = null;
-function getGeminiClient$1() {
-  if (geminiClient$1) return geminiClient$1;
-  const project = process.env.GOOGLE_CLOUD_PROJECT;
-  const location = process.env.GOOGLE_CLOUD_LOCATION ?? "us-central1";
-  if (!project) {
-    throw new Error("GOOGLE_CLOUD_PROJECT environment variable is required");
-  }
-  geminiClient$1 = new GoogleGenAI({ vertexai: true, project, location });
-  return geminiClient$1;
-}
 let openaiClient$1 = null;
 function getOpenAIClient$1() {
   if (openaiClient$1) return openaiClient$1;
@@ -429,7 +438,7 @@ async function getWhisperInstance() {
   return whisperInitPromise;
 }
 async function transcribeGemini(audioBuffer) {
-  const client = getGeminiClient$1();
+  const client = getGeminiClient();
   const base64Audio = audioBuffer.toString("base64");
   const response = await client.models.generateContent({
     model: "gemini-2.5-flash",
@@ -491,17 +500,6 @@ async function transcribe(audioData, provider = "local") {
   logger.info({ provider, text }, "Transcribed");
   return text;
 }
-let geminiClient = null;
-function getGeminiClient() {
-  if (geminiClient) return geminiClient;
-  const project = process.env.GOOGLE_CLOUD_PROJECT;
-  const location = process.env.GOOGLE_CLOUD_LOCATION ?? "us-central1";
-  if (!project) {
-    throw new Error("GOOGLE_CLOUD_PROJECT environment variable is required");
-  }
-  geminiClient = new GoogleGenAI({ vertexai: true, project, location });
-  return geminiClient;
-}
 let openaiClient = null;
 function getOpenAIClient() {
   if (openaiClient) return openaiClient;

package/package.json

@@ -1,9 +1,13 @@
 {
   "name": "pi-voice",
-  "version": "0.2.0",
+  "version": "0.3.3",
   "description": "Voice interface for pi coding agent",
   "author": "Yuku Kotani",
   "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/yukukotani/pi-voice"
+  },
   "type": "module",
   "main": "./out/main/index.js",
   "bin": {