PixCull

PixCull is an open-source culling and review workflow for photographers and video creators who want granular judgments, export formats, and the option to keep processing on their own hardware. It is best suited to technically confident users managing sizeable shoots, rather than anyone seeking a ready-made mobile app.

The software’s breadth is impressive for a free MIT-licensed project, but its cloud-judging default and do-it-yourself iOS companion deserve a deliberate setup before it becomes part of a professional workflow.

Overview

PixCull sorts photographs into keep, maybe, and cull, then scores every frame from one to five stars for technical quality, subject, composition, light, moment, and aesthetic. A written explanation accompanies both the overall decision and each scoring axis. That makes close calls more auditable than a single opaque ranking, although reviewing six dimensions can add time for photographers who only want a fast reject pass.

The scope extends to video. PixCull can rate clips, identify potential reel material, transcribe audio, and create an EDL for Adobe Premiere or DaVinci Resolve. The MIT license permits commercial use and forking, making the project workable for studios that want to adapt or self-host their tooling.

Key features

Photo analysis

RAW support, subject detection, duplicate detection, blur detection, and closed-eye detection address the practical reasons a frame is rejected. PixCull describes profiles for nine genres, including wedding, wildlife, sports, landscape, portrait, event, journalism, commercial, and still-life photography. The genre range is useful for mixed practices, while the per-axis rationale gives experienced shooters material to challenge or refine in their own review process.

Exports and connected workflows

XMP sidecars can carry selections into Lightroom and Capture One. IPTC captions and standalone HTML galleries cover metadata delivery and simple presentation without requiring a separate gallery service. PixCull can watch a Lightroom or Capture One tether destination, and folder-based synchronization works through iCloud, Dropbox, or a NAS, which suits studios moving files between workstations.

Local, cloud, and mobile operation

RAW decoding, scoring, face analysis, and GPS analysis run locally, so images can remain on the user’s machine. However, the maker’s security policy says photos are sent to MiniMax for judging by default from version 2.50 after recorded consent. The --vlm-mode off option disables that path. With cached models and cloud judging disabled, a run can operate without a network connection. There is no telemetry, and the local server binds to 127.0.0.1 by default; LAN deployment can be protected with an API-key header.

The iOS swipe companion is distributed as a Swift package to build in Xcode, not through the App Store. That is a useful route for developers or organizations able to maintain a build, but a significant barrier for photographers expecting an install-and-go phone companion.

Pricing

PixCull — free. The MIT-licensed open-source software has no subscription charge. The maker notes an optional paid licence, while optional cloud model calls may create separate API costs. That combination favors users who can run local models and manage their own infrastructure; teams depending on cloud judging should budget beyond the headline zero price.

Platforms

PixCull supports API, iOS, Linux, macOS, self-hosted deployments, and Windows. The iOS offering requires an Xcode build, while self-hosting and local-server controls make the desktop and server options more suitable for technically managed environments than casual users.

Who it's for

Wedding and event photographers can use the keep/maybe/cull workflow, closed-eye and blur checks, tether watching, and Lightroom or Capture One sidecars to shorten large deliveries. Sports, wildlife, and landscape shooters benefit from genre coverage, duplicate detection, and detailed moment and composition scores. Photojournalists and studios gain IPTC captions, HTML galleries, self-hosting, and commercial-use licensing. Self-taught photographers may appreciate the written rationales, but should expect more configuration than a conventional consumer app.

Pros and cons

Pros

• Six-axis, explained scoring: separates technical, subject, composition, light, moment, and aesthetic judgments instead of hiding them in one number.
• Broad delivery options: XMP, IPTC, HTML galleries, and video EDL export connect selections to established editing and publishing tools.
• Local control: local decoding and analysis, no telemetry, offline operation with cached models, and a localhost-bound server support privacy-conscious workflows.
• Open licensing: MIT terms allow commercial use and forking.

Cons

• Cloud judging is the default: users handling sensitive imagery must actively disable MiniMax processing with --vlm-mode off.
• Mobile setup is demanding: the iOS companion is a Swift package for Xcode, not an App Store download.
• Operational responsibility: only the latest main branch is supported, with no backported patches for older commits; vulnerability reports go to the maintainer.
• Potential usage costs: optional cloud model calls can add API charges despite the free software price.

Alternatives

AI Photo Culling Software is the broader category to consult when you want to compare culling tools rather than commit to PixCull’s open-source workflow.

Zno is a freemium option with Android, iOS, macOS, web, and Windows coverage, plus a free start and paid instant-gallery packages; choose it when gallery-oriented, multi-device access matters more than self-hosting.

Facet is free and local on Linux, macOS, self-hosted, web, and Windows, with no cloud, accounts, or subscriptions; pick it when avoiding cloud processing is the overriding requirement.

Project Kestrel offers a free open-source desktop app for Linux, macOS, and Windows, with cloud compute available; consider it when you want a simpler desktop entry point.

Optyx provides free previews, image rating, metadata, and Lightroom support on macOS and Windows, with a $6.99-per-month annual-billing Pro plan; choose it when a commercial Lightroom-focused product is preferable.

RawPicker is a freemium culling alternative spanning Android, iOS, Linux, macOS, web, and Windows.

Kepla is a macOS-focused paid product with a free private preview; consider it when you want a single-platform workflow.

ShotSieve is free and open source on Linux, macOS, and Windows under AGPLv3+, with separate terms for learned-IQA dependencies and model weights.

Futura Photo offers a free Windows Community plan with unlimited analyses of up to 100 images per analysis and one concurrent installation, plus paid tiers; pick it when a packaged Windows service and explicit analysis caps fit better.

Verdict

Choose PixCull if you want unusually detailed, explainable photo and video culling, strong Lightroom/Capture One interchange, and the freedom to run an MIT-licensed workflow locally or self-hosted. Look elsewhere if you need an App Store mobile companion, a cloud-free default, or vendor-maintained support across older builds without configuration work.