Security & Privacy Center

A plain-language description of how DisputeDocs Pro actually handles credit reports and personal information. We describe only what the product does today. We do not claim any certification we have not obtained: DisputeDocs Pro is not SOC 2, ISO 27001, or PCI certified, and no independent penetration test has been performed to date.

• Credit reports, bureau responses and evidence go to private storage buckets. There is no public URL.
• Access is granted only through short-lived signed URLs issued after a server-side authorization check.
• Uploads are validated by file type, magic bytes and size before they are accepted.
• Every read and deletion of a stored document is written to an access log with the actor and reason.
• Files are scoped to the owning workspace and client record; they are deleted when the record is deleted.

Extracted data is minimized before it is ever written to the database: Social Security numbers are reduced to the last four digits, account and card numbers are masked, dates of birth are reduced to a birth year, and driver's license or government ID numbers are stripped. Free text such as remarks and finding descriptions is redacted with the same rules.

To be accurate about the document itself: the original PDF you upload is retained in private storage as you uploaded it. If that PDF contains a full SSN, that copy contains it too. You may redact the document before uploading, and you may delete the file at any time from the report page.

• AI is used for four purposes: report extraction, error-pattern detection, letter polishing, and reading bureau response letters.
• Extraction requires the report document itself; the other three receive only the specific fields needed for that task.
• Raw consumer credit files are never used to train models, ours or a provider's, and are never used for marketing or advertising profiles.
• Model output cannot state a legal conclusion. Wording that asserts a violation, an entitlement to damages, or a guaranteed outcome is rewritten before you see it.
• Every AI process is documented in our internal processing register with its purpose, data categories, retention and human-review posture.

The app refuses to assist CPNs or credit privacy numbers, new credit identities, file segregation, using an EIN in place of an SSN, false identity-theft claims, fabricated police reports or statutes, blanket "dispute everything" demands, and any misrepresentation to a bureau or furnisher. These refusals are enforced in code and covered by an automated safety test suite that runs on every build.

Every customer record carries the workspace that owns it, and row-level security policies restrict reads and writes to members of that workspace. Role permissions (owner, admin, manager, compliance reviewer, agent, viewer) are enforced on the server and in the database, not only in the interface. An automated regression suite re-checks these policies after every database change.

• Consumers can export only their own data; firms can export only data for clients they are authorized for.
• Account deletion is requested in-app, with a confirmation step and visible status tracking.
• Deletion runs as tracked jobs per artifact: database records, stored documents, AI artifacts and analytics.
• Encrypted backups age out on their own schedule; the backup expiry date is recorded on the deletion job.

We use a managed cloud database and storage provider, a payment processor, an AI gateway, and transactional email. Categories, roles and instructions are described in our Data Processing Addendum. We do not sell personal information.

Email privacy@disputedocs.app with a description and reproduction steps. We acknowledge reports within three business days and do not pursue good-faith researchers.