Imadeo
Self-hosted · open source

Your photos.
Your server.

Imadeo is a home for everything you have ever shot — organised by date, by who is in it, and by what it shows. It runs on hardware you own, and your library never leaves it.

Imadeo Search your photos and videos… Upload EA
Photos 248 items Sort by Date taken

Aug 5, 2026

Nov 21, 2025

2026
2025
2024
2023
2020

Everything a library needs

Built around the way people actually look for a picture: roughly when it happened, who was there, or what was in the frame.

A timeline that scrolls decades

Photos group by the day they were taken, with a year rail down the side so a jump to 2013 is one click, not a minute of scrolling.

People, found automatically

Faces are detected and clustered as photos arrive. Name someone once and every picture of them is grouped — merge, split, or correct a wrong match at any time.

Search by what's in the picture

Describe the shot — "a dog on a beach" — and get results without a single tag or filename. Filenames, dates, albums and folders are searchable too.

Duplicates worth reviewing

Perceptual hashing finds re-saved, resized and re-sent copies of the same shot — not just identical files — and groups them so you choose what to keep.

Albums and real folders

Upload a folder and its structure comes across intact. Albums sit alongside, so curated sets and the original tree can both exist.

Accounts you control

Invite family by email, or create local accounts directly. A locked area keeps private photos out of the main timeline.

No one else is in the loop

Every part of Imadeo runs on your machine — including the parts that recognise faces and understand what a photo shows.

0

Photos uploaded to anyone else's cloud

On-device

Face and image recognition run locally — no external API, no per-photo cost

Yours

Files stay in a plain folder you can back up, move, or walk away with

Download

Imadeo runs wherever Docker does — a NAS, a home server, or the machine already under your desk.

Docker

The full stack — app, database, queue and the machine-learning worker — in one compose file. Works on Synology, Unraid, TrueNAS and any Linux box.

Available now Installation guide

TrueNAS SCALE

Install from the shell with the compose file, or wait for the catalog app to land under Apps → Discover with the library dataset picked from a form.

Available now Installation guide