iGemini

A powerful Apple Shortcut to seamlessly integrate Google Gemini across your entire Apple ecosystem: iOS, iPadOS, macOS, and watchOS. Bring the advanced AI capabilities of Gemini to your daily workflows, activating it from Siri, the Share Sheet (on iPhone/iPad), or directly from your Apple Watch.

Overview

iGemini bridges the gap between Google Gemini and Apple's native ecosystem, allowing you to interact with Gemini naturally from anywhere on your device — no app switching or browser tabs required.
It acts as a lightweight interface that sends your queries to Gemini and returns intelligent, contextual answers right inside the Apple Shortcuts environment.

Installation

If you haven't created an API key yet, get one here:
https://aistudio.google.com/api-keys

Then, download the shortcut from the official repository:

Once downloaded:

1. Tap on the downloaded file — it will automatically open the Shortcuts app on your device.
2. Tap “Add Shortcut” to import it.
3. Enter your Gemini API key and select your preferred language when prompted (en, it, de, es).

NOTE: Language codes are CASE SENSITIVE. Please use lowercase: en for English, it for Italian, de for German, or es for Spanish.

4. iGemini is now ready to use — you can invoke it from Siri, the Share Sheet, or your Apple Watch.

Support my work

If you want to support my projects, consider buying me a coffee!

Usage

iGemini can be triggered in several ways — from Siri, the Share Sheet, Apple Watch, or directly inside the Shortcuts app.
It automatically detects the type of content you’re sharing (text, link, webpage, PDF, or image) and adapts the prompt sent to Gemini accordingly.

---

Quick Start

1. Run the shortcut or say “Hey Siri, Gemini”.
2. Type or speak your question.
3. Gemini will generate a response instantly.
4. Choose what to do with the result:
   • Copy to clipboard
   • Continue the conversation
   • Create a new note
   • Chose link from a list provided from Gemini

---

Feature Highlight — “Capture the Screen” Mode
If your input includes phrases such as “capture the screen”, “take a screenshot”, or similar, iGemini automatically takes a real-time screenshot of your device’s display, that will be sent to Gemini together with your text prompt, embedding visual analysis of your screen.

You can use it to ask context-aware questions like:

“Take a screenshot and explain this error.”
“Capture the screen and summarize what’s visible.”
“Take a screenshot and describe the app layout.”

From the Share Sheet

You can invoke iGemini directly from the Share menu in Safari, Notes, Files, or Photos.
Depending on the type of content, the shortcut behaves differently:

Input Type	What Happens
Text	Gemini receives your selected text and answers your question about it.
Link / URL	The linked page is fetched, cleaned, and analyzed by Gemini.
Web Page	The full webpage content is extracted and summarized or explained on demand.
PDF	The text content is extracted from the PDF and sent to Gemini for analysis or summarization.
Image	The image is converted to Base64 and described or analyzed by Gemini using multimodal input.

When the share sheet opens, simply tap Gemini, type your question (e.g., “Summarize this article”), and wait for the reply.

---

Conversational Mode

After receiving a response, you can tap Continue Chat to ask follow-up questions.
Gemini keeps the context of the current session locally, allowing for natural back-and-forth interaction.

Example:

You: “Summarize this article.”
Gemini: “It discusses the impact of AI on education…”
You: “Now give me three key takeaways.”

The previous context is preserved automatically — no need to re-share the same content.

---

On Apple Watch

• Open iGemini from the Shortcuts app, complication or Siri.
• Speak or type your question directly on your watch.

---

On macOS

Run the shortcut from the Shortcuts menu bar or assign it to a keyboard shortcut.

---

Technical Flow

1. User Input
   The shortcut starts by asking for user input through a text field or dictated speech.
   The text you enter becomes the message sent to Gemini.

2. Language and Model Setup
   On the first run, you are asked to choose a preferred language (e.g., en, it).
   This choice determines which internal prompt structure is used (en_str, it_str, etc.).
   The model used by default is gemini-flash-latest, but you can edit the shortcut to change it.

3. Input Type Recognition

Before building the request, iGemini detects the kind of content you provided and adapts the payload accordingly (most detections are triggered when the Shortcut is invoked from the Share Sheet):

• Text — Default: the text you enter or paste in the Shortcut is used as the prompt for Gemini’s text generation.
• PDF — The Shortcut extracts the PDF text (using Shortcuts’ “Get Text from PDF”/PDF parsing actions) and sends that extracted text as the content to be analyzed or summarized.
• Web page / URL — When a web page or URL is shared, iGemini attempts to fetch and extract the page content (or uses the URL as a reference) before sending it to Gemini. The exact result depends on the source and what Shortcuts can retrieve from that page.
• Image — Images are encoded in Base64 and included in the request payload (inline data) so Gemini can analyze visual content.

This logic ensures the request payload matches the input type: extracted text for PDFs/pages, Base64-encoded data for images/files, and plain prompt text for simple text input.

4. Request Construction
   iGemini builds a JSON payload following Gemini’s API structure:

   • "contents" contains your message text.
   • The API key and model name are inserted into the request URL:
     https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent?key={apiKey}

5. Sending the Request
   The Shortcut uses the Get Contents of URL action to send the POST request to Gemini’s endpoint.
   The headers include Content-Type: application/json.

6. Response Handling
   The response from Gemini is decoded from JSON.
   The shortcut extracts the text output from the "candidates[0].content.parts[0].text" field.

7. Output Presentation The result is shown using a menu with multiple options:

   • Show Answer – Displays Gemini’s full response in a text view.
   • Continue Chat – Reopens the input field, keeping the previous conversation context.
   • Copy Response – Copies Gemini’s message to the clipboard.
   • View Links – If Gemini’s response includes any URLs, they are automatically detected and listed for direct access.
   • Exit – Ends the session without saving.

8. Update Check (optional)
   On each run, iGemini compares its internal version number with the one available online
   at https://raw.githubusercontent.com/gabrielerandazzo/iGemini/main/version.json.
   If a newer version is found, it offers an update prompt.

---

Example Prompts

• “Summarize this PDF in three bullet points.”
• “Explain this webpage as if I were a beginner.”
• “Describe what’s in this image.”
• “Translate this text into Italian.”
• “Generate five video ideas based on this paragraph.”

---

Updating iGemini

The shortcut periodically checks for updates from the repository.
If a new version is available, you’ll be prompted to download it automatically.

---

Notes & Limitations

• The API key you enter is stored directly inside the Shortcut’s variables — it is not encrypted or securely stored. Anyone with access to your Shortcuts data can view it, so handle it carefully.
• The project is officially named iGemini, but the actual Shortcut file is named Gemini — this choice improves Siri recognition when invoking it by voice.

Legal Notice

This project is an independent, unofficial Shortcut. It is not affiliated with, endorsed by, or supported by Apple Inc. or Google LLC. “Gemini” is a trademark of Google LLC. All trademarks and product names mentioned belong to their respective owners.

Contributors

Created by Gabriele Randazzo

Want to help improve this project?
All contributions are welcome — bug fixes, feature ideas, and documentation improvements.

Every contributor will be credited here in future releases.

Name	Contribution
Gabriele Randazzo	Creator & Maintainer
-	-

If you’d like to contribute:

1. Fork the repository
2. Make your changes
3. Open a pull request