fltman

fltman / gemma4-mac

Public

Run Gemma 4 e4b locally on Apple Silicon — text, vision, and Apple Photos auto-captioning

19
2
100% credibility
Found Apr 28, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A macOS tool for Apple Silicon that runs a local AI model to chat about text/images/audio, auto-generate captions and keywords for selected Apple Photos, and curate balanced yearbook albums using photo metadata.

How It Works

1
🔍 Discover photo AI magic

You hear about a handy tool that adds smart captions and organizes your Apple Photos using your Mac's own power, all running locally.

2
📥 Download and set up

Grab the files and run the simple setup script—it checks your Mac, prepares everything, and adds easy shortcuts to your terminal.

3
🖥️ Open a fresh terminal

Start a new chat window on your Mac, and the AI is ready to go after downloading its brain once.

4
💬 Chat or describe anything

Ask questions, describe pictures or sounds you point to—it feels like having a clever friend right on your Mac.

5
Pick your photo adventure
✏️
Caption selected photos

In Photos app, pick pictures, run the magic command, and watch captions and tags appear.

📚
Create a yearbook album

Tell it a year or date range, and it smartly picks the best shots to make a new album.

🎉 Your photos shine

Sit back with captioned memories, organized albums, and no cloud needed—everything private on your Mac.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is gemma4-mac?

gemma4-mac lets you run Gemma 4 e4b locally on Apple Silicon Macs like Mac Mini M4, Mac Studio, or MacBook Pro M5, handling text chat, image analysis, and audio transcription entirely on-device with Python and MLX. Fire up the `gemma` CLI for interactive REPL sessions, one-shot prompts like "describe this photo," or multimodal inputs with multiple images and audio files. It solves the hassle of cloud dependency for Gemma 4 on macOS by adding seamless Apple Photos integration: auto-caption selected photos in Swedish (or custom styles), merge keywords without overwriting, and curate yearbook albums from date ranges.

Why is it gaining traction?

Unlike Ollama's simpler text focus, this delivers vision and audio natively via MLX for blazing speeds—28 tokens/sec text, 57 for images on M5—plus unique Photos tricks like pulling iCloud previews for captioning without downloads. Devs dig the install script that sets up a venv, downloads the 4-bit model, and adds shell aliases, with dry-run previews and clean uninstall. Running Gemma 4 on Mac Mini or MacBook means no API keys, just local privacy for scripting AI around your photo library.

Who should use this?

Apple Silicon owners building local AI tools for macOS apps or personal workflows, especially those with massive Photos libraries needing quick captions, tags, or auto-curated "best of 2024" albums. Photo hobbyists or family archivists on MacBook Pro M5 who want Gemma 4's vision smarts to enhance metadata without exporting files. Script kiddies automating Gemma 4 on Mac for batch image descriptions or event clustering.

Verdict

Grab it if you're on Apple Silicon and want Gemma 4 on macOS with Photos superpowers—docs are crisp, perf is solid, and setup takes minutes despite 19 stars and 1.0% credibility signaling early days. Skip for production until more testing; it's a promising playground for local Gemma 4 mlx experiments.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.