Project Kestrel

Overview

Project Kestrel is a desktop tool for sorting, reviewing, and searching bird photographs. It is best suited to bird photographers who want to compare burst shots and export ratings to an editor; its North American species classification and optional cloud tools are important boundaries to weigh.

Its strongest case is a focused workflow that combines scene grouping, bird-aware image scoring, and species search while keeping desktop work local. Worldwide photographers can still tag species manually, but should not expect automatic classification beyond North American birds.

Key features

Culling and search

Kestrel groups high-speed bursts into scenes, making it easier to compare near-duplicate frames rather than review a whole outing as an undifferentiated set. Its detection model identifies and segments birds before scoring, so background pixels do not influence quality ratings. A separate machine-learning score weighs noise, motion blur, and focus, then sorts sharper frames first. That combination is relevant to photographers handling many similar shots, though the scoring is a way to prioritize review, not a substitute for choosing the keeper.

Detected birds can be tagged by species and family, allowing searches across folders and outings. Automatic classification currently covers North American species; elsewhere, manual tagging is the route to searchable species labels. Kestrel exports ratings to Lightroom, Darktable, Capture One, and other editors, which makes it more useful alongside an existing editing workflow than as a replacement for one.

Local work and sharing

The desktop app works locally without an account or internet connection, and photos remain on the machine unless a cloud or sharing feature is used. Perch is the sharing option: users can upload an outing, annotate its timeline, and share it. Cloud Compute handles images for analysis; source images are deleted after analysis, with cleanup and storage-expiry safeguards if deletion fails. Results are deleted when downloaded or after 30 days if left undownloaded, while job records keep counts and timestamps for billing. Users who want to avoid uploading photos can stay with the local app and skip those optional services.

Pricing

The desktop app is free and open source. Windows and macOS downloads are provided; Linux users run it from source. The separate Apple App Store version costs 10.00 USD once and is the same app as the free direct download, with automatic updates included. It suits Mac users who prefer the App Store update route; the direct download avoids that purchase.

Cloud Compute has a free allowance of 2,500 images for every new account, with no card required. For larger cloud workloads, Blue Jay costs $20 a month for 15,000 images per month, while Sandpiper costs $50 a month for 50,000 images per month. The paid tiers fit users whose image volumes exceed the free allowance; photographers who do not need cloud analysis can use the local desktop app instead.

Platforms

Kestrel supports Linux, macOS, and Windows, although Linux users run the desktop app from source rather than downloading it as a packaged build. The App Store purchase is for macOS. Perch and Cloud Compute are optional services, not requirements for local desktop use.

Who it's for

Kestrel is a good fit for bird photographers who shoot bursts, want bird-specific culling signals, and organize outings by species. Its export options suit people who already finish images in Lightroom, Darktable, Capture One, or another editor. Automatic species tagging is most useful for North American collections; photographers focused elsewhere will need to tag manually. The maker describes it as a one-person project built for the developer's own bird photography, a useful context for readers considering how specialized its focus is.

Pros and cons

• Pros: Bird segmentation keeps backgrounds out of quality scoring, while burst grouping and sharpness ranking help narrow similar frames.
• Pros: Local use needs neither an account nor internet, and photos stay on the device unless optional cloud or sharing features are used.
• Pros: Ratings export to several established editors, so culling can feed into an existing workflow.
• Cons: Automatic species classification is limited to North American birds; other regions require manual tags.
• Cons: Linux requires running from source, which is less direct than the Windows and macOS downloads.
• Cons: Cloud capacity beyond the 2,500-image new-account allowance requires a monthly plan, making the larger-volume workflow a recurring cost.

Alternatives

For a broader set of culling choices, browse AI Photo Culling Software. Zno is worth considering for photographers who want Android, iOS, web, and desktop access alongside a free starting plan and free trial. Choose Facet if a free local tool with no cloud, accounts, or subscriptions is the priority. PixCull offers a free, MIT-licensed open-source option, with optional cloud model calls potentially adding API costs. Optyx is another Lightroom-compatible choice, with a free tier for previews, image rating, and metadata.

ShotSieve is a free and open-source option for Linux, macOS, and Windows. AccioPix Photo Culling is a free web tool with no sign-up and no photo limit beyond the computer's capacity. Futura Photo offers a Windows Community tier with unlimited analyses of up to 100 images each. Kepla is another option to consider.

Verdict

Choose Project Kestrel if bird photography is your main use case and you want burst-aware culling, bird-focused scoring, species search, and editor rating exports in a local desktop workflow. Look elsewhere if automatic species classification outside North America is essential, or if you need Linux without running software from source.