EFFECTIVE 16 SEPTEMBER 2026
Privacy
Data Preflight is designed to inspect supported CSV files without sending them to us. Paid data, automation, and software projects follow a separate written scope and data-handling agreement.
Local CSV tools
Files selected in the free scanner are read and analysed by code in your browser. Safe repair previews and repaired CSV files are also generated on your device. The import rehearsal compares a current export with a planned import on the same device. File contents, generated reports, repaired files, and row-level rehearsal results are not uploaded to Data Preflight, and we do not receive or retain them. Each selected CSV can contain up to 10 MB and 100,000 data rows.
Import rehearsal can optionally remember risk thresholds and up to five recent summary records in your browser storage. This setting is off until you enable it. The saved summary contains the date, platform, import-mode assumption, and aggregate result counts. It does not contain filenames, headers, cell values, SKUs, product names, or row-level changes. Turning the setting off or using its clear action removes this local history from the browser.
Project brief builder
You can submit a project brief through the online form or use the email-draft fallback. Online submission sends the fields you enter, including your reply-to email, name if provided, project categories, timing, budget range, and description, to Data Preflight. The form does not accept file uploads. Submitted briefs are stored in a Cloudflare D1 database and are visible only in the authenticated administration view for assessing and managing the request.
Opening an email draft, copying the brief, or copying our address does not submit the brief to us. You must review and send the draft in your mail app. Do not enter customer records, passwords, payment data, private links, or other sensitive information in either channel. You may request deletion of a submitted enquiry by emailing us with its reference number.
If you separately opt in to an AI-assisted initial assessment, we may send only the text description and project category, format, volume, target date, and requested deliverable to Cloudflare Workers AI. We do not include your name or reply-to email in the model request. The model selects possible scope questions from a fixed list; it does not set a binding price. A preliminary reply may be sent from our Zoho Mail account when that feature is enabled. A person must review and confirm any formal proposal. Your choice is recorded with the enquiry. Without your opt-in, the brief is not passed to the AI assessment service.
Basic hosting and anonymous analytics
Cloudflare may process ordinary technical information needed to deliver, protect, and measure the site, such as an IP address, request time, requested path, browser information, performance data, and security events. We use Cloudflare Web Analytics for aggregate traffic and page performance measurement.
The site also sends a small set of anonymous product events, such as whether a scan started or completed, the selected platform and interface language, broad file-size and row-count ranges, the result category, issue codes and counts, template and report downloads, repaired-file downloads, completed import rehearsals, rehearsal report downloads, repair-brief copies, project-enquiry or product-feedback draft opens, copied contact addresses, and the selected service or broad format, technology, or target category. An AI-assistant referral visit is counted only when the browser provides a recognized referring site; only a fixed assistant name is sent, never the full referring URL or search question. Online enquiry contents and contact details are stored separately from these anonymous product events. An email-draft event records only the action; it cannot tell us whether a message was sent, received, or replied to. These events never include an email address entered by a visitor, email subject or body, file name, CSV contents, column names, cell values, SKUs, product names, selected repair rules, repair counts, rehearsal thresholds, import modes, result rows, report contents, repaired-file contents, or project-brief text. We do not use these events for advertising or cross-site tracking.
Paid service projects
Before receiving a working file, the written project scope should define:
- the purpose, file fields, and processing instructions;
- the approved transfer and storage method;
- who may access the working data;
- whether any subprocessors or AI services are permitted; and
- the delivery, retention, and deletion timetable.
Additional terms, including a data processing agreement or cross-border transfer clauses, may be required for a particular project.
Your choices
You can use the free scanner without an account. You can also decline a proposed file-transfer method or request a different retention period before paid work begins. Privacy requests for a service project should be sent to contact@data-preflight.com or through the verified marketplace channel used for that engagement.
This page describes the current website behaviour. A signed project agreement controls if its data-handling terms are more specific.