Open source

Cull your photos in minutes, not hours.

Selecta groups your shots by scene, discards the blurry ones, and picks the best of each group, telling you why. It runs on your computer and your original photos never leave it.

Download it now — free on GitHub
✦Free & open source — bring your own API key✦
Selecta — Review by groups Analyzing scenes…2 picks · 4 discarded
Scene 1 · 4 shots1 kept
Motion blur
AI pick
Duplicate
›Sharpest frame, subject well placed, no clipped highlights.
Scene 2 · 3 shots1 kept
Overexposed
AI pick
+ more
›Eyes in focus and a cleaner background than the rest of the burst.
Illustrative interface · sample numbers only A accept · X reject — all from the keyboard
Scroll
// The problem

After every shoot, hundreds of near-identical frames.

Bursts, small framing changes, a few soft shots: choosing the best ones is the most tedious part of photography.

Professional culling tools are built for working photographers, and they cost money. Hobbyists are left without a simple option focused on their own genre.

// How it works

Local first, AI only where it matters.

  1. 01

    Pick a folder

    Point Selecta at a folder of JPEGs, choose the genre and how strict it should be.

  2. 02

    Local analysis

    Sharpness, motion blur, exposure and near-duplicates, checked on your PC for free.

  3. 03

    Scene grouping

    Similar shots are gathered into scenes using capture time and visual similarity.

  4. 04

    AI picks the best

    Only small previews of the candidates go to the model you choose, with a one-line reason.

  5. 05

    Review & export

    Correct anything with a key press, export the picks to a subfolder. Originals untouched.

// Features

Built for photographers, not for pipelines.

Genre profiles

Aviation, portrait, landscape, car and urban each get their own evaluation criteria.

Strictness levels

Keep everything but the obvious rejects, go balanced, or keep only the best. Thresholds are visible.

Explains every decision

Each pick comes with a short reason, so it also helps you sharpen your own eye.

Cost under control

See an estimate before the analysis starts and a live token counter while it runs.

Learns your taste Planned

Your corrections refine the next selections.

XMP export Planned

Ratings and labels you can read in Lightroom-style tools.

// Video guides

Get your API key, then let Selecta work.

Short videos that show where to find your key for each provider and how to use it in Selecta.

Your first selection

  1. Download Selecta from GitHub and start it.
  2. Open Settings → Personal Cloud API and paste your key, or pick Ollama.
  3. Choose a folder, the genre and the strictness, then start the analysis.
  4. Review the picks with A / X and export them.
github.com/LeoFrizzy28/Selecta-V0.1

Your key is saved only on your PC. Never share it with anyone.

// Who it's for

For people who shoot a lot and keep a little.

Hobbyist photographers

Aviation and spotting, nature, urban, events: anyone home with hundreds of shots.

Fast pre-selection

Work with JPEGs and want a first pass before opening Lightroom or similar tools.

Privacy-minded users

Prefer not to upload a whole shoot to a cloud service.

Developers

Curious people who want to read the code and contribute to an open source project.

// Privacy & your own key

Your originals stay on your computer.

Only 1024 px previews go to the AI provider you choose, using your own API key. No account, no telemetry, and the local server listens only on your machine.

Bring your own model
Google Gemini OpenAI Anthropic Ollama (local)

Previews are handled under the terms of the provider you pick. Entry-level models cost roughly $0.20 per 1,000 photos (estimate; prices change).

Follow the project, or help build it.

Star it, try it, open an issue. Selecta is open source and growing one phase at a time.

View on GitHub