simoncirstoiu

simoncirstoiu / alice

Public

Analyse · Learn · Ingest · Curate · Export — AI-powered YOLO dataset management toolkit

ai ai-tools annotation computer-vision dataset
85
6
100% credibility
Found Apr 20, 2026 at 85 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

ALICE is a web app for browsing camera snapshots, drawing object outlines, cleaning datasets, and training custom vision models from your own photos.

How It Works

1
📥 Download and launch Alice

You grab the simple tool and run it to open a friendly web app in your browser.

2
📁 Set up your folders

You point Alice to where your camera snapshots and photo collections are stored so it can find them easily.

3
🤖 Pick a smart vision helper

You choose and download a ready vision model that knows common objects like people or cars.

4
👆 Mark objects in snapshots

You browse your camera photos, draw colorful boxes around people, pets, or vehicles with easy drag-and-drop tools.

5
🧹 Spot and clean duplicates

Alice finds repeated photos and overlapping marks, letting you tidy up your collection quickly.

6
🚀 Train your custom vision

You run the step-by-step trainer, watching it learn from your marked photos to build a personalized smart detector.

🎉 Your detector is ready!

Alice hands you a tuned vision model perfect for your cameras, spotting exactly what you need.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 85 to 85 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 alice?

Alice (simoncirstoiu/alice on GitHub) is a self-contained Python web app for YOLO dataset workflows: browse/annotate images, ingest Frigate snapshots or video frames, run AI detection, dedupe via perceptual hash, and train models end-to-end. Fire up `./alice.py` for a localhost:8080 UI with canvas editing, gallery view, stats, and a 5-step trainer (export-dedup-annotate-train-ONNX). Solves scattered-tool pain for custom object detection on home cams, like analyse GitHub repo footage into labelled datasets.

Why is it gaining traction?

85 stars signal Frigate+YOLO niche love: one-file build (python3 builder.py), Docker GPU passthrough, pHash dupes across cams, auto-merge AI boxes (IoU>0.5), real-time train metrics. Beats LabelStudio/LabelImg by bundling viewer+trainer—no cloud, minimal deps (ultralytics optional via UI), keyboard-driven like alice keyboard GitHub tools.

Who should use this?

Frigate NVR owners training person/car models on event clips; security cam tinkerers dodging Roboflow fees; AI devs analysing GitHub repository AI datasets for quick YOLOv8/11 fine-tunes.

Verdict

Strong pick for Frigate stacks—mature UI/docs punch above 85 stars and 1.0% credibility (early but battle-tested). CC BY-NC limits commercial; test on small datasets first.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.