Data Subject Access Requests under GDPR require organizations to disclose all personal data they hold about an individual. Before those documents can be released, every piece of third-party personal information, names, email addresses, phone numbers, addresses, national insurance numbers, bank details, must be permanently redacted. This page explains how SafeRedact automates that redaction step, what it costs, how it compares to alternatives, and when you should use it.
The scale problem with DSAR redaction
DSAR volume keeps rising. Every major privacy regulation, UK GDPR, EU GDPR, CCPA, and the growing patchwork of US state privacy laws, grants individuals the right to access their personal data, and public awareness of that right continues to grow. Regulatory enforcement has sharpened too: the ICO has taken action against organisations, including government departments, that repeatedly missed the legal deadline for answering subject access requests.
A single DSAR can involve 20,000 or more files. When a data subject requests everything an organization holds about them, the response package typically includes emails (EML and MSG files), Microsoft Teams chat transcripts exported from Purview eDiscovery, spreadsheets with mixed data in individual cells, PDF reports, Word documents, and every attachment nested inside those emails. One person, often a DPO or privacy analyst, is typically responsible for processing the entire package, with a one-month statutory deadline under GDPR, extendable where a request is complex, or a 45-day deadline under CCPA.
The file types are the hard part. EML files contain nested quoted replies where third-party PII appears across multiple layers of conversation history. MSG files are binary and require parsing before content is even readable. HTML exports from Microsoft Teams and Purview eDiscovery have their own structure. CSV files contain mixed data types in single cells: a single cell might include a name, address, and phone number together. Each format requires different extraction and detection logic. Manual review of 20,000 files at this level of complexity is not feasible within any reasonable deadline.
Why visual masking is not redaction
PDF overlay redaction, placing black boxes over sensitive text, is not permanent redaction. The text content underneath the visual overlay remains in the file and can be recovered using copy-paste, standard text extraction tools, or by removing the annotation layer in any PDF editor. ICO enforcement actions have cited recoverable redaction as a compliance failure, and several high-profile data breaches have resulted from organizations releasing documents with visual-only redaction that recipients were able to reverse. If the underlying text is still present in the file, it has not been redacted.
Permanent redaction requires removing the text content itself, not masking it visually. SafeRedact produces plaintext output with [REDACTED_TYPE] markers, for example, [REDACTED_NAME], [REDACTED_EMAIL], [REDACTED_NINO]. The original PII is deleted from the output, not hidden behind a visual element. There is no annotation layer to remove, no hidden text to extract, and no metadata to recover. The redaction is irreversible by design.
How SafeRedact works
1. Upload your eDiscovery export
Drag and drop a Purview export, a PST, a ZIP of individual files, or select files manually. SafeRedact processes PDF, DOCX, XLSX, TXT, HTML (including Teams transcripts), CSV and JSON, and email as PST, EML or MSG; the file types page has the full list and the edges. Nested ZIPs are extracted automatically. Files stay in the browser: only extracted text and the case's data subject details are sent for AI detection.
2. AI detection across all files
SafeRedact uses a two-layer detection pipeline. Layer 1 is deterministic regex matching for structured identifiers: National Insurance numbers, email addresses, phone numbers, postcodes, sort codes, and bank account numbers. Layer 2 is contextual AI analysis using Anthropic's Claude, which identifies names, addresses, salaries, dates of birth, and other PII that requires understanding context. In DSAR mode, the pipeline preserves the data subject's information and redacts everyone else's. Processing runs unattended in the background; a large batch can run overnight.
3. Review every detection
Every AI detection is presented for human review in a split-pane interface. The left panel shows the original document with PII highlighted: yellow for third-party data that will be redacted, green for the data subject's data that will be preserved. The right panel shows the redacted output with [REDACTED_TYPE] markers. Click any detection to see its type, confidence score, and source (regex or AI). Accept or reject each detection individually. Add manual redactions for anything the AI missed by selecting text and choosing a PII type. When you are satisfied with every file, attest under your own account that review is complete.
4. Export the reviewed case
Download the complete package: the disclosure pack of redacted documents, and an internal audit folder with the audit trail, the review decision log, the detection report, the subject preservation summary and a manifest of any files that failed to process. If you paid for the case at set-up, nothing is charged here. On the free evaluation, this is where you pay for the case.
Priced per case, before you start
Most redaction tools charge a subscription before you see results. SafeRedact charges per case: the price is set by the number of documents in the case and shown at set-up, before processing begins, and it does not change as you add to that case within its band; a case that reaches the ceiling of its band moves to the next band for the difference in price. Your first case is refunded in full within 14 days of export if you ask, whatever the reason.
Evaluate first: process and review up to 500 of your own documents, free, for 14 days. Pay for the case when you download the redacted pack.
How SafeRedact compares
| Feature | SafeRedact | Manual Process |
|---|---|---|
| Handles eDiscovery exports (PST, EML, MSG, Teams) | ✓ | ✗ |
| AI PII detection | ✓ | ✗ |
| Manual redaction override | ✓ | ✓ |
| Reviewer sign-off with audit trail | ✓ | ✗ |
| Price known before processing | ✓ | N/A |
| Per-job pricing (no subscription required) | ✓ | N/A |
| DPA included | ✓ | N/A |
Privacy management platforms such as OneTrust and Osano manage the request workflow: intake, identity verification, tracking and deadlines. SafeRedact is a dedicated tool for the document redaction step and can be used alongside them, between document collection and disclosure.
Pricing
Each case is priced by document count. The price is shown at set-up and paid by card before processing starts, or you can evaluate first and pay when you download. Cases above 25,000 documents are quoted in writing. There is no subscription and no minimum commitment. See how pricing works or contact us to discuss a case.
Your first case, step by step →
Compliance infrastructure
No document content is stored on SafeRedact servers. What is retained for each file is a content-free processing record: a case identifier, a file index, a user identifier, character and byte counts, and a one-way fingerprint of the extracted text. SafeRedact's sub-processors, listed in the sub-processor schedule, each hold their own SOC 2 Type II report. See the trust page for detail.
The Data Processing Agreement for your framework is accepted on screen when you set up your organisation, and a signed copy is emailed to you. The UK GDPR, EU GDPR and CCPA agreements are published at /enterprise/legal.
Frequently asked questions
What file types does SafeRedact support for DSAR redaction?
SafeRedact processes PDF, DOCX, XLSX, TXT, HTML (including Microsoft Teams transcripts exported from Purview eDiscovery), CSV and JSON, and email as PST, EML or MSG. Purview exports are supported directly: drag and drop the exported ZIP or PST and SafeRedact unpacks it and routes every file by type. A scanned, image-only PDF yields no text to scan and is flagged during review; standalone images are not processed. The file types page has the full list.
How long does a 20,000-file DSAR take to process?
Processing runs unattended in the background and the browser tab does not need to stay open. Processing time follows how much text the documents hold, not how many there are: measured on real cases, short emails and documents of a page or two process at about 1,300 to 2,000 an hour, and a mailbox export with typical attachments at about 70 to 110 documents an hour. The timeline table gives the figures for 2,500 and 10,000 documents, and the pre-flight screen estimates the time for your own files before processing starts.
What is the difference between SafeRedact and OneTrust for DSARs?
OneTrust manages the DSAR workflow: intake, identity verification, tracking and deadline management. SafeRedact is a dedicated tool for the document redaction step: identifying and removing third-party personal data from the files before you disclose them to the data subject. The two can be used together.
Is there a Data Processing Agreement?
Yes. When you set up your organisation you accept SafeRedact's Data Processing Agreement for your framework (UK GDPR, EU GDPR or CCPA) on screen, and a signed copy with a certificate of execution is emailed to you. The agreements are published at /enterprise/legal for review before you sign up.
Does SafeRedact store document content?
No. Documents stay in the browser: only extracted text and the case's data subject details are sent for AI detection. SafeRedact keeps only account records and a content-free processing record per file (case identifier, file index, user identifier, character and byte counts, and a one-way fingerprint of the extracted text), with no file names or document content. Text sent for detection is never used for model training and is deleted by the AI provider within 30 days.
When do I pay?
You can start without paying. The free evaluation lets you process and review up to 500 of your own documents for 14 days, with the full review, and you pay for the case when you download the redacted pack. Otherwise you pay for a case at set-up, before processing starts, at the price shown for its document count. Either way you upload files, run AI detection, review every detection in the split-pane interface, add manual redactions for anything the AI missed, and attest under your own account that review is complete; export is open once you have confirmed the redacted output is ready for disclosure. Organisations billed by invoice pay on the terms agreed with them.