Add:Added Gemini AI Support

README.md

@@ -6,10 +6,12 @@
 
 <br />
 
-**Potomatic** is a command-line tool for translating `.pot` (Portable Object Template) files into multiple languages using AI (currently OpenAI). We built it to streamline large-scale localization of WordPress products, a process we detail in [this behind‑the‑scenes article](https://www.gravitykit.com/translating-wordpress-plugins-using-chatgpt/).
+**Potomatic** is a command-line tool for translating `.pot` (Portable Object Template) files into multiple languages using AI. We built it to streamline large-scale localization of WordPress products, a process we detail in [this behind‑the‑scenes article](https://www.gravitykit.com/translating-wordpress-plugins-using-chatgpt/).
 
 While [`gpt-po`](https://github.com/ryanhex53/gpt-po) helped us get started, we needed smarter retry logic, cost controls, and better visibility into large jobs, among other things. **Potomatic** delivers those improvements and more, as well adds fine‑grained prompt tuning through a built‑in [A/B testing utility](#-ab-testing-for-prompt-optimization).
 
+Supports multiple AI providers: **OpenAI**, **Google Gemini**, and **Google Translate API**.
+
 ## 📢 Disclaimer
 
 Translation quality varies depending on factors such as model selection, prompt design, and the complexity of the source text. **Potomatic** can generate a baseline translation, but the output should always be reviewed and verified before use.
@@ -40,7 +42,7 @@ For improved results, consider refining your prompt, using a higher-tier model,
 
 ## 🚀 Key Features
 
-* **🤖 AI‑powered translations** – Translate into any language supported by OpenAI models.
+* **🤖 AI‑powered translations** – Translate into any language supported by OpenAI, Google Gemini, or Google Translate API.
 * **📦 Smart batch handling** – Tune batch size, concurrency and retries for the right balance of cost and speed.
 * **💰 Cost‑conscious execution** – Accurately estimate costs and tokens, and control the maximum cost of a job.
 * **🔄 Incremental & resumable workflows** – Resume interrupted jobs, merge with existing `.po` files, or force a re‑translation.
@@ -285,6 +287,73 @@ Using this example, "Block Editor" and other terms will not be translated to tar
 
 ---
 
+## 🤖 AI Providers
+
+Potomatic supports multiple AI providers for translation:
+
+### Google Gemini
+
+Uses Google's Gemini models for translation. Set the provider to `gemini` and configure your API key.
+
+**Available Models:**
+
+| Model | Prompt Cost | Completion Cost | Best For |
+|-------|-------------|-----------------|----------|
+| `gemini-3.1-pro-preview` | $0.00125/1K tokens | $0.005/1K tokens | Complex translations, highest quality |
+| `gemini-2.5-pro` | $0.0005/1K tokens | $0.0015/1K tokens | High-quality general use |
+| `gemini-2.5-flash` | $0.000175/1K tokens | $0.000525/1K tokens | Fast, cost-effective translations |
+| `gemini-flash-latest` | $0.000175/1K tokens | $0.000525/1K tokens | Latest optimized flash model |
+| `gemini-3.1-flash-lite-preview` | $0.0000375/1K tokens | $0.00015/1K tokens | Ultra-budget translations |
+
+**Usage:**
+```bash
+# Using Gemini 2.5 Flash
+./potomatic -l fr_FR -p translations.pot --provider gemini --model gemini-2.5-flash -k $GOOGLE_API_KEY
+
+# Using the latest flash model
+./potomatic -l fr_FR -p translations.pot --provider gemini --model gemini-flash-latest -k $GOOGLE_API_KEY
+
+# Using budget-friendly flash-lite
+./potomatic -l fr_FR -p translations.pot --provider gemini --model gemini-3.1-flash-lite-preview -k $GOOGLE_API_KEY
+```
+
+### Google Translate API
+
+Uses Google Cloud Translate API for fast, cost-effective translations. Ideal for straightforward text without complex nuances.
+
+**Pricing:** Google Translate API charges based on character count:
+- $20 per 1M characters (standard tier)
+- Volume discounts available for higher usage
+
+**Usage:**
+```bash
+./potomatic -l fr_FR -p translations.pot --provider google-translate -k $GOOGLE_CLOUD_API_KEY
+```
+
+### OpenAI
+
+Uses OpenAI's GPT models for translation. This was the original provider and offers the widest model selection.
+
+**Available Models:**
+
+| Model | Prompt Cost | Completion Cost | Best For |
+|-------|-------------|-----------------|----------|
+| `gpt-4o-mini` | $0.00015/1K tokens | $0.0006/1K tokens | Fast, budget-friendly |
+| `gpt-4o` | $0.0025/1K tokens | $0.01/1K tokens | High-quality |
+| `gpt-4o-mini-2024-07-18` | $0.00015/1K tokens | $0.0006/1K tokens | Specific dated version |
+
+**Usage:**
+```bash
+./potomatic -l fr_FR -p translations.pot --provider openai --model gpt-4o-mini -k $OPENAI_API_KEY
+```
+
+**Environment Variables:**
+- `GOOGLE_API_KEY` - For Gemini provider
+- `GOOGLE_CLOUD_API_KEY` - For Google Translate provider
+- `API_KEY` - Falls back to OpenAI
+
+---
+
 ## ⚙️ Configuration Files
 
 **Potomatic** uses several configuration files in the `config/` directory to customize its behavior:

config/gemini-pricing.json

@@ -0,0 +1,28 @@
+{
+  "models": {
+    "gemini-2.5-pro": {
+      "prompt": 0.0005,
+      "completion": 0.0015
+    },
+    "gemini-2.5-flash": {
+      "prompt": 0.000175,
+      "completion": 0.000525
+    },
+    "gemini-3.1-pro-preview": {
+      "prompt": 0.00125,
+      "completion": 0.005
+    },
+    "gemini-flash-latest": {
+      "prompt": 0.000175,
+      "completion": 0.000525
+    },
+    "gemini-3.1-flash-lite-preview": {
+      "prompt": 0.0000375,
+      "completion": 0.00015
+    }
+  },
+  "fallback": {
+    "prompt": 0.0005,
+    "completion": 0.0015
+  }
+}

config/google-translate-pricing.json

@@ -0,0 +1,12 @@
+{
+    "models": {
+        "default": {
+            "prompt": 0,
+            "completion": 0
+        }
+    },
+    "fallback": {
+        "prompt": 0,
+        "completion": 0
+    }
+}

package.json

@@ -32,6 +32,7 @@
 	},
 	"scripts": {
 		"translate": "./potomatic",
+		"translate:gemini": "./potomatic --provider gemini",
 		"ab-prompt-test": "node tools/ab-prompt-test",
 		"test": "vitest run",
 		"test:watch": "vitest",
@@ -45,6 +46,7 @@
 		"node": ">=18"
 	},
 	"dependencies": {
+		"@google/generative-ai": "^0.24.1",
 		"chalk": "^5.3.0",
 		"commander": "^12.0.0",
 		"dotenv": "^16.5.0",

src/config/index.js

@@ -216,7 +216,8 @@ export function parseCliArguments() {
 		.option('--locale-format <format>', 'Format to use for locale codes in file names: `wp_locale` (ru_RU), `iso_639_1` (ru), `iso_639_2` (rus), or `target_lang` (default)', DEFAULTS.LOCALE_FORMAT)
 
 		// === Translation Options ==.=
-		.option('-k, --api-key <key>', 'OpenAI API key (overrides API_KEY env var)')
+		.option('--provider <provider>', 'AI provider to use (e.g., "openai", "gemini", "google-translate")', DEFAULTS.PROVIDER)
+		.option('-k, --api-key <key>', 'Provider API key (overrides API_KEY env var)')
 		.option('-m, --model <model>', 'AI model name (e.g., "gpt-4o-mini")', DEFAULTS.MODEL)
 		.option('--temperature <number>', 'Creativity level (0.0-2.0); lower = more deterministic, higher = more creative', (val) => Math.max(0, Math.min(2, parseFloat(val))), DEFAULTS.TEMPERATURE)
 		.option('-F, --force-translate', 'Re-translate all strings, ignoring any existing translations', DEFAULTS.FORCE_TRANSLATE)
@@ -302,7 +303,15 @@ export function validateConfiguration(options) {
 	const errors = [];
 
 	if (!options.dryRun && !options.apiKey && !process.env.API_KEY) {
-		errors.push('🔑 API key required (set API_KEY env var, use --api-key, or try --dry-run)');
+		// Google Translate doesn't require an API key
+		if (options.provider !== 'google-translate') {
+			errors.push('🔑 API key required (set API_KEY env var, use --api-key, or try --dry-run)');
+		}
+	}
+
+	// Google Translate requires source language
+	if (options.provider === 'google-translate' && !options.sourceLanguage && !process.env.SOURCE_LANGUAGE) {
+		errors.push('🌐 Source language required for Google Translate (use -s or --source-language)');
 	}
 
 	if (!options.targetLanguages || options.targetLanguages.length === 0) {