🎨 EDICO

Eliz & Delia's Image Captioner and Organizer — the free, open-source way to caption and curate a consistent character dataset for LoRA training.

EDICO turns a messy folder of character images into a clean, structured, AI-captioned dataset — in one beautiful browser app. No Python, no Node, no command-line wizardry. Just pick a model, drop in your images, and let EDICO organize the rest.

✨ Why EDICO?

🌐 One app, not two worlds.
The whole workflow lives in one polished browser interface — no terminal hopping between setup and review.

🏷️ Structured captions.
Pose, gaze, framing, lighting — consistent keyword captions instead of freeform text soup.

🎛️ A real visual organizer.
Filter, edit, rearrange, auto-arrange, and re-sequence your dataset with drag-and-drop ease.

🤍 Free & private.
MIT licensed. Runs 100% locally through Ollama — your images never leave your machine.

🔬 The workflow

👋

Welcome

Step 1

Open an existing project or start fresh — progress is always saved.

🧠

Vision Model

Step 2

EDICO detects your Ollama models and flags the vision-capable ones with [VL].

🪪

Identity

Step 3

Set your trigger word and the character's fixed traits.

🖼️

Images

Step 4

Drag & drop a folder of images. Clean thumbnails, remove on the fly.

🏷️

Pipeline

Step 6

Live captioning with per-image status, cancel, and resume.

🗂️

Organizer

Step 7

Filter, edit, add, rearrange, and export your finished dataset.

🖥️ Built for creators

Local & private — runs on your own machine through Ollama.
No Python, no Node — a single PowerShell file + your browser.
Structured caption editor — click terms, never hand-write raw text files.
Smart badges — spot duplicate captions and non-standard resolutions instantly.
Rearrange & auto-arrange — drag to order, or sort by up to three priorities.
Resume anytime — interrupted runs pick up where they left off.
Open & MIT-licensed — free forever, yours to read and modify.

🚀 Try it in minutes

Windows

start.bat

Your browser opens EDICO automatically — no setup script, no venv, no dependencies beyond Ollama.

📚 Learn more

User Guide & Docs 🐙 GitHub