FREE BETA
Terms of use, help and limitations
PhotosClaires checks an already extracted Google Photos Takeout export and can create a new organized copy, entirely on your computer.
Terms of use for the beta
By using PhotosClaires, you agree to these terms. Access is free and requires no purchase, subscription or provision of personal data for commercial purposes.
You must use the tool only on files that you are authorised to access and in compliance with applicable law. You retain ownership of your files and remain responsible for your backups and for any decisions made after reading the report.
You must not deliberately disrupt the service, circumvent its security measures or use it to infringe the rights of others. Access may be restricted in the event of abuse or technical necessity.
The beta is experimental. To the extent permitted by law, the publisher does not guarantee exhaustive results, continuous availability or compatibility with every export variant. Nothing in these terms deprives a user of any mandatory rights that apply to them.
These terms are governed by French law. They may change as the product evolves; the date shown at the bottom of the page identifies the applicable version.
What the beta does
- it inventories recognized media and JSON files;
- it attempts to match each JSON file to its media without hiding ambiguous matches;
- it separately counts available photo-taken dates and
creationTimefallback dates; - it checks some exact duplicates with a SHA-256 hash;
- it produces a downloadable aggregate report;
- with your approval, it can copy all recognized media to a new session folder;
- it places missing, contradictory or ambiguous dates in an
_A-verifierfolder and creates a CSV manifest.
The source export is never renamed, moved, corrected or deleted. The organized copy keeps all recognized media, including exact duplicates, uses a random session name and rejects any destination it detects as already existing.
Current technical limitations
- a maximum of 100,000 files per selection;
- JSON files larger than 8 MB are skipped and reported;
- JSON reading stops after 256 MB per scan;
- media larger than 256 MB is not loaded into memory for duplicate detection;
- duplicate checking reads at most 10,000 media files, 4 GB of content and 2,000 groups of files with identical sizes;
- duplicates are searched for only among files of the same size and refer exclusively to byte-for-byte identical content;
- similar photos, bursts, edited versions and visually similar videos are not treated as duplicates;
- unusual Takeout names, incomplete exports and future changes to Google's format may result in orphaned JSON files or ambiguous matches;
- direct copying requires a recent desktop version of Chrome or Edge and explicit permission to write files;
- the browser cannot guarantee the destination's available space: check that the stated amount is free;
- the copy organizes only high-confidence photo-taken dates;
creationTimedates, medium-confidence matches, conflicts and missing dates remain in_A-verifier; - the copy rewrites neither EXIF/QuickTime fields nor file-system dates;
- a session is complete only when
COMPLETE.txtis present andINCOMPLETE.txtis absent; a very early failure may prevent the first marker from being created.
Dates and how to interpret the report
A date found in photoTakenTime is counted as an available photo-taken date. If it is missing, a valid creationTime date is counted separately as a fallback. The latter may describe creation or processing within the service and does not necessarily prove the actual capture date.
Only photoTakenTime dates from high-confidence matches determine the year/month folder in the new copy, in UTC. Items with only creationTime remain in a subfolder for review. The results remain technical indicators: check the warnings, the manifest and a sample of files before making any decision in another application.
Compatibility and good practice
Selecting a complete folder and writing the copy rely on browser features primarily supported by recent desktop versions of Chrome and Edge. First extract every Takeout archive into a local folder.
- Keep an untouched copy of the original export.
- Close applications that use a large amount of memory.
- If the export exceeds the limit, analyze its parts separately.
- Do not delete any file solely on the basis of the beta report.
- Choose a destination with at least the stated amount of free space and keep the source folder until you have fully verified the result.
- Choose a parent folder that is different from the source Takeout export.
- On Safari or Firefox, download the CSV plan; direct copying is not offered.
Availability and support
The service is provided free of charge as a beta, with no guarantee of continuous availability or compatibility with every export. It may change, be interrupted or display incomplete results; warnings are part of the scan report.
Support and beta feedback: greg2ajaccio@gmail.com. Describe your browser and operating system, the approximate number of files, whether the scan and copy completed, and your main difficulty. Never send media, JSON files, file paths or screenshots containing file names.