Export formats

1. What this page is

Article 26 of the EU Data Act asks a cloud service to keep an online register of the data it can export, its structure and its formats, so that a customer can leave without asking permission and without surprises. This is that register. Everything below can be produced by the customer alone, from the app, at no charge — see section 13 of the terms of service.

2. Your files: ZIP archives

• What is inside. The original files exactly as they were uploaded — bytes untouched, never re-encoded, whatever the format (JPEG, PNG, HEIC, WebP, MP4, MOV, PDF and so on). Where a tape was uploaded as a browser-made preview first and the original followed, the original is what you get; where only the preview was ever uploaded, that is what exists. On the casting accounts, every photo and video is replaced by its compressed copy once that copy has passed its checks, so the compressed copy is the file you get.
• Structure. One archive per folder, per selection or per project section, as requested. File names are the uploaded names; when two files in one archive share a name, the second becomes name-2.ext, the third name-3.ext, so nothing is silently overwritten.
• Format. Standard ZIP with ZIP64 extensions, entries stored uncompressed (photos and video are already compressed), so any operating system opens it and extraction runs at disk speed. An archive holds up to 5,000 files; for a bigger set, download one folder at a time.
• Not inside. The thumbnails and preview sizes the gallery displays. They are generated from the originals and can be regenerated by anyone from the files above.

3. Feedback and metadata: CSV, XLSX, JSON

Three formats carry the same information; pick the one your tools read.

CSV — UTF-8, one header row, fields quoted per RFC 4180, CRLF line ends. One row per vote, frame vote, like, note or comment, plus one row per folder without feedback so nothing is missing. Columns, in order:

section, name, ref, status, recommended, type, frame, person, login, audience, decision_maker, vote, first_choice, text, timecode, at followed by the card fields of a casting (age, height, agency and so on); a location scout has none.

XLSX — an Office Open XML workbook made to be shared, with two sheets: Summary, every folder someone liked, voted on or commented on, with its cover image and one column per person showing their vote and like; and People, each person's likes, votes and comments, and whether they finished. Every individual row, with its time, is in the CSV and JSON.

Google Sheets — the same workbook, sent from your browser to your own Google Drive, where Google turns it into a Sheet you can edit and share. It is a copy taken at the moment you export and does not update. Export again for a fresh one; the earlier Sheet stays as you left it. The permission Google asks you for only lets us create that file, and no Google token reaches our servers (privacy policy).

JSON — a single UTF-8 document, timestamps in ISO 8601 with time zone, with these top-level keys:

• exported_at, project (id, slug, name, client name, kind, status, dates);
• rounds — the review rounds and their state;
• people — every person who voted or commented: display name, login, audience, decision-maker flag, the email they gave, last seen;
• folders — one object per folder: its card fields, reference code, tally, and the arrays votes, frame_votes, likes, notes, comments (each with person, value, text, timecode where relevant, and time);
• media — one object per file: id, folder, file name, type, size, dimensions or duration, upload time;
• submissions — each time a client sent a round: who, when, message;
• downloads — see the download log below.

4. Reports: PDF

The Selects report is a laid-out document for people, per audience (client-facing or internal): the same folders, votes and notes as above, with images. It is not a data format and not needed to move to another service; the JSON and the ZIP together contain everything the report shows.

5. Download log

Who downloaded what: a JSON list, also on the project's Feedback page, with the time, the person and login, their audience, the scope (folder, selection or project), the status, the number of files and bytes, and anything that was skipped.

6. What is not exported

Data specific to the internal working of the service, which Article 25 of the Data Act lets a provider keep out of an export and which is of no use elsewhere: web server logs, the job queue, derived thumbnails and preview sizes, session tokens and password hashes. Client logins are exported by name and login, never with a password; the people who use them set new ones wherever you go next.

7. How to get them

• In the app: open a project, then Export. CSV, XLSX and JSON download at once; a PDF report and large ZIP archives are prepared as jobs and appear in the same screen with a download link when ready. Folder and selection ZIPs are also available to your clients in the gallery when you allow downloads.
• Through the API, with a staff session: GET /api/projects/<id>/export.csv, …/export.json, …/export.xlsx, …/report.pdf, …/download-log; POST /api/projects/<id>/exports queues a large export and GET /api/exports/<id>/file fetches it.
• Whole workspace: write to support@preproselects.com and we prepare every project's archives and data files together, at no charge, within the transition period the terms describe.

Last updated 23 October 2026. See also the terms of service and the legal notice.