# eleyed > Zero-retention PII redaction for text and PDFs, including scanned PDFs and photographed documents saved as PDF. Built-in OCR — no separate conversion step. Detects names, contacts, financial identifiers (NPI), health identifiers (PHI), and payment card data, plus caller-supplied terms and regex. Redaction is permanent — content is removed from the file, not covered over. Nothing you upload is stored: no data at rest, no third-party AI. Browser, REST API, and an MCP server so an AI agent can scrub a document before it reaches a model. Credit-based pricing; no subscription required. ## Product - [Home](https://eleyed.com) — What eleyed does, who uses it (legal, HR, healthcare, financial services, developers), how it works in three steps. - [Pricing](https://eleyed.com/pricing) — Credit packs and monthly subscription tiers. All prices exclude tax; tax is handled by Onelink under Stripe Managed Payments (Merchant of Record). - [Free PII checker](https://eleyed.com/pii-checker) — Paste text and see what personal data it contains, and get it back redacted for up to 3,000 characters. No account, nothing stored. - [Remove personal data before you paste into AI](https://eleyed.com/remove-personal-data-before-ai) — Strip names, contact details and ID numbers from text before it goes into ChatGPT, Claude, Gemini or Copilot. Free up to 3,000 characters, no account. - [SAR deadline calculator](https://eleyed.com/sar-deadline-calculator) — When a UK GDPR subject access response is due: one month, extensions, clarification pauses, and bank holidays for England and Wales, Scotland and Northern Ireland. Runs in the browser. - [FAQ](https://eleyed.com/faq) — Questions on pricing, data handling, refund policy, supported file formats, and NPI/PHI/PCI coverage. - [Terms of Service](https://eleyed.com/terms) — Full terms including data handling and refund policy. EU/UK 14-day statutory withdrawal right applies to initial subscription payments. - [Guides](https://eleyed.com/guides) — Evergreen reference on redaction practice: - [SAR and DSAR redaction](https://eleyed.com/guides/dsar-redaction) - [FOI redaction](https://eleyed.com/guides/foi-redaction) - [Choosing a redaction tool](https://eleyed.com/guides/choosing-a-redaction-tool) - [Why a black box isn't redaction](https://eleyed.com/guides/black-box-is-not-redaction) - [What still leaks after you redact a PDF](https://eleyed.com/guides/pdf-hidden-content) - [How to redact a PDF](https://eleyed.com/guides/how-to-redact-a-pdf) - [How to redact a bank statement](https://eleyed.com/guides/redact-bank-statement) ## What eleyed detects - Personal identity: names, email addresses, phone numbers, physical addresses, dates of birth - Government IDs: US SSN, driver licence, passport, UK NHS numbers, UK National Insurance numbers - Financial (NPI under GLBA): credit card numbers (also PCI cardholder data), bank account numbers (IBAN, US), cryptocurrency wallet addresses - Healthcare (PHI under HIPAA): medical licence numbers, plus standard identity fields where they qualify as PHI in a clinical context - Network and web: IP addresses, URLs - Custom terms and regex patterns supplied by the caller ## How eleyed works - **Scanned PDFs are supported, including photographed pages and faxes saved as PDF.** Built-in OCR reads the words on the page, so no separate conversion step is needed. - **Use it in the browser, call the REST API, or connect the MCP server** from an agent. Same detection, same retention model, whichever way it is reached. - **Redaction is permanent.** Anything found is taken off the page itself, rather than covered with a box that can be lifted away later — the failure that has caused real disclosure incidents elsewhere. - **Deep Analysis** takes a second look at the finished page and picks up anything the first read struggled with, such as words set inside a logo or letterhead. - **Searchable output** keeps the finished PDF searchable by keyword, while leaving everything redacted out of it. - **Nothing is stored.** Documents never reach a disk. Files passing through the service exist only in memory, and are discarded as soon as the work is done. There is no document store. - **No third-party AI.** All detection runs on eleyed's own infrastructure. Customer content is never sent to OpenAI, Anthropic, Google or any other outside provider — the requirement that rules most redaction tools out of legal and healthcare work. ## Scope - eleyed detects and redacts the data types listed above. The GLBA / HIPAA / PCI-DSS operational, contractual and audit obligations remain with the organisation using the service. - Payments are processed by Onelink under Stripe Managed Payments, which handles VAT and sales tax across jurisdictions. ## For developers - Browser and REST API. Two endpoints: `POST /api/v1/redact/text` and `POST /api/v1/redact/pdf`. Authenticated via `X-API-Key` header. Full reference (parameters, response shapes, error codes, entity types) available in the customer dashboard after signup. Same pipeline as the web app, including OCR on PDFs. - **MCP server** — [overview and install snippet](https://eleyed.com/mcp). Endpoint `POST /api/mcp`, Streamable HTTP, authenticated with the same `X-API-Key` header. Registered as `com.eleyed/redact`. Tools: `redact_text` (strip personal data from a string before it is sent to a model) and `list_entity_types` (free). Text only — PDF redaction is REST-only, because a file does not fit in a tool call. Credits, rate limits and the zero-retention guarantee are identical to the REST API. ## Contact - Support: support@eleyed.com - Payments Merchant of Record: Onelink (via Stripe Managed Payments) - Operator and data controller details are on the [Terms of Service](https://eleyed.com/terms) page.