Privacy Policy

Accessible Content Preflight respects your privacy. This starter policy describes how we intend to handle information as the platform evolves.

What runs locally

Deterministic caption and alt-text checks, WCAG contrast measurement, screen reader experience preview (text), and platform checklists run in your browser. They are not sent to our server for processing.

Caption, working caption, alt text, and platform selection may be stored in your browser's session storage so they persist when you move between platform checker pages in the same tab. Images are not stored in session storage.

Optional AI image analysis

If you choose AI image analysis, your image, caption, alt text, and platform selection are transmitted to this application's server and then to OpenAI for processing. This application does not intentionally persist uploaded images or AI inputs in a database or on disk.

Video transcription and visual review (Phase 6)

When you generate transcripts and captions, your video is uploaded to private temporary storage on our server infrastructure, then sent to OpenAI for transcription. The application attempts to delete the temporary video after processing. Orphan cleanup may remove files older than a configured retention window.

Sampled video frames are sent to OpenAI only when you explicitly request visual information review. Frames are not intentionally stored by this application.

Transcripts and caption cues remain in your browser session unless you download them. This application does not intentionally persist them in a database.

Screen Reader Experience Preview — speech

Optional speech playback uses your browser's speech synthesis API. Behavior depends on your browser and operating system. Speech text is processed locally and is not sent to our server.

Batch accessibility checker (Phase 8)

CSV and ZIP imports are parsed in your browser. Batch rows and analysis results are kept in memory for the current session only (not intentionally stored in a database or on our servers). Only images you explicitly select for AI review are sent through the same secure image analysis route as the single-post checker.

Website accessibility scanner (Phase 9)

When you submit a public website URL, our server-side scanner visits that page in an isolated browser, renders it, and runs automated accessibility rules. We do not intentionally store full page HTML. Scan results remain in your current browser session unless you export them. We do not collect authentication credentials for website scanning.

If you explicitly request AI fix guidance, short rule summaries and sanitized HTML snippets from findings may be sent to OpenAI. Scanning itself does not require OpenAI.

Social, image, contrast, video, and batch tools offer optional AI-Assisted Remediation through the same shared API when you click generate — for example caption revisions, alt options, contrast design guidance, or a single caption cue. Original content and measured contrast values stay in the UI until you choose to apply a suggestion.

If you choose View on Page, we re-visit the same public URL on demand, highlight the affected element in a viewport screenshot, and return the image to your browser for this session only. We do not store those screenshots on our servers or use OpenAI for visual location.

Copy Developer Ticket builds remediation handoff text from your current scan results entirely in your browser. Copying or downloading a ticket does not send it to Jira, GitHub, or other services, and does not use OpenAI unless you separately chose to include optional cached AI Fix Lab guidance in the ticket preview.

Accessibility Preflight Report (Phase 7)

Reports are assembled in your browser from session data (caption checks, contrast samples, optional AI results already obtained, screen reader preview, platform confirmations, and video caption state). Export actions do not send your draft to our servers for report generation. Downloaded files remain on your device; we do not host share links or report history.

Accessibility blog (editorial)

Public guides on /blog are stored in a separate MongoDB database when configured. That database holds editorial articles only (titles, Markdown content, SEO metadata, authors). It does not store your scan results, uploaded media, captions, tickets, or other tool session data. Reading the blog does not require using the accessibility checkers.

Platform checklists

Platform publishing guidance does not connect to your social media accounts. Before-you-publish confirmations are kept in your browser session only.

Third-party processing

OpenAI processes content you submit through the AI feature according to its own terms and policies. Do not upload sensitive or confidential content unless you accept that third-party processing risk.

Storage approach

We aim to process AI requests temporarily without intentional long-term storage of your uploaded content in this application.

Contact

Questions about privacy may be sent to support@contentpreflight.example.com.