All-in-one redaction means a single platform redacts personal data from video, audio, documents and screen recordings in one workflow. Redaction is the permanent removal or masking of personal data from a file before it is shared. One records request can bring CCTV footage, a call recording and a folder of PDFs. A team can end up with a tool for video and another for audio. Add a PDF editor for paperwork and a spreadsheet to track what went where.
That works until a single request mixes file types. Then the gaps start to cost time, because a file that moves through a second tool needs a second review pass. A reviewer repeats the same third-party decision in three tools, or a missed name forces a second pass before the deadline.
Why do separate redaction tools leave gaps?
Picture a council handling a subject access request under Article 15 of the GDPR. A subject access request is the right to see the personal data an organization holds about you. The requester wants everything held about them. The reply might include 47 minutes of CCTV from a leisure center and a 9 minute recorded call to the contact line. It might also hold a 12 page complaint file. Most items hold other people's personal data, including a bystander's face in the footage and a neighbor's address in a letter.
With three tools, staff learn three interfaces and check three sets of export settings. Nobody holds a single record of who redacted what, or when.
The gaps show up at review. A face is blurred in the video, but the same person's name survives in the audio. A PDF looks redacted, yet the text sits underneath the black box and can be copied out. Each tool did its job. Nothing checked the whole response.
Time is the other cost. A reviewer decides that the requester's partner counts as a third party. That call must then be made again in each tool, and a different reviewer can reach a different answer. The deadline does not move while they do. Under UK GDPR a subject access request must be answered within one month. The deadline can be extended by up to two further months, but only for complex or numerous requests. The ICO's right of access guidance explains the rules.
What does all-in-one redaction include?
An all-in-one redaction platform handles personal data across file types in one workspace. In Secure Redact that covers recorded video, audio, documents and screen recordings. It uses a single login and shared permissions, and every file goes through the same review process.
Video
Pimloc's models detect faces, number plates and other identifiers in recorded footage at 720p or above, then blur them. Reviewers can correct what the models miss before export. Our video redaction tool works on CCTV, body-worn and dashcam files.
Detection accuracy depends on the footage, and we have not published a single detection rate for that reason. On video below 720p, in poor light, or with small distant subjects, the tool misses faces it would catch on clear footage. The gap widens the further the subject is from the camera. Treat any clip in those conditions as needing a full manual review, not a spot check. To get a rate for your own cameras, count the faces missed in a ten minute sample of your footage. A person has to watch the output.
Audio
Recordings carry spoken names, addresses, dates of birth and card numbers. Audio redaction software finds these and lets you bleep or remove them. Accuracy drops on any section where two people speak at once, so treat every overlapping section as unchecked until a reviewer has listened to it. The transcript of overlapping speech is less reliable, and a name in the quieter voice can be missed. It also drops when background noise is close to the volume of the speech. Phone recordings saved below 32 kbps are heavily compressed, so listen to them from start to finish. We do not publish one figure for the drop, because it varies by recording. To measure it, take ten minutes of your own calls, listen for every name and compare that list with what the tool flagged. Reviewers should listen to every flagged section and the passages around it.
Documents and screen recordings
Document redaction covers PDFs and case files. A proper redaction removes the underlying text from the file. Drawing a black rectangle over it does not. You can test any tool on this point by exporting a file and then trying to copy the covered text.
Screen recordings need the same care. A recording of a claims system or a support session shows names and account numbers. A notification can flash up for less than a second, which is between 1 and 24 frames at 25 frames per second. A reviewer watching at normal playback speed can miss one, so step through screen recordings frame by frame. Automated detection scans frames for text. Test it with a clip that holds a one second notification. It can still miss very small text or text with weak contrast against its background.
File types at a glance
The table sets out what each file type involves and where a reviewer should look hardest.
| File type | Personal data handled | What a reviewer checks |
|---|---|---|
| Video | Faces, number plates, on screen text | Missed detections in low light or at distance |
| Audio | Spoken names, addresses, account numbers | Crosstalk and background noise |
| Documents | Names, IDs, addresses, financial details | Text removed from the file, not just covered |
| Screen recordings | Names and records shown on screen | Notifications that appear briefly |
What does a shared workflow add?
One queue shows every item in a request. Reviewers get role-based permissions, so one person can redact and another approves.
That also keeps decisions consistent. When a request names one individual, you want the same calls applied everywhere. One reviewer sees the video, the call and the file together, and decides once who is the requester and who is a third party.
Evidence needs a record. A chain of custody log is a record of each action on a file, who took it and when. If a requester or a court challenges a disclosure, the log shows which redactions were applied and who approved them. That matters in a criminal case, where the Criminal Procedure and Investigations Act 1996 requires prosecutors to disclose unused material to the defense. If the defense challenges a redacted file, you need to show it was not altered beyond the agreed redactions.
There is a security gain too. Every export from one tool to another is another copy of sensitive material to track and delete. Moving a request through three tools can leave three copies of each file, while one platform keeps a single working copy plus the redacted output.
Teams with their own case management system can connect it through a redaction API. Before you plan an integration, ask the vendor which file formats, file size limits and setup steps the API supports.
Where do teams use all-in-one redaction?
A US police records unit answering a FOIA request has to withhold personal privacy information under exemptions 6 and 7(C). The official FOIA portal sets out the request process. Federal agencies have 20 working days to make a determination, and state deadlines differ. A request about one traffic stop can cover bodycam video, the 911 call audio and the incident report.
Patient requests to a hospital bring a similar mix. Video from a ward corridor and scanned notes can both identify other patients. The HIPAA Privacy Rule applies to protected health information whatever the format.
Insurers can face the same bundle in a single motor claim, which might include dashcam footage and a recorded phone statement. In the UK, a retailer or school faces the same one month deadline when a requester asks for CCTV footage and call recordings.
What are the limits of all-in-one redaction?
Automatic detection does not replace review. No model finds every face or every spoken name. Plan a human check on every export that leaves your organization, and budget reviewer time for poor quality files.
A single platform also means a single vendor. If a specialist tool beats it on one file type, you accept that trade in return for one workflow. A team that only ever redacts video gains little from it.
Software cannot decide what to redact either. Your policy does that. Whether a bystander's face is exempt, or a staff member's name must stay, is a legal call.
How do you evaluate an all-in-one redaction platform?
Run the same checks on every vendor you shortlist.
Use your own files. Test a real, messy sample, such as a night shift CCTV clip, a noisy call, a scanned PDF and a screen recording.
Check the export. Try to recover covered text from documents and blurred faces from video.
Ask for the audit log. It should show who did what, and when, for every file type.
Review permissions. Redactors and approvers should be separate roles.
Confirm storage and access. Find out where files are held and who can reach them, then compare that with your own policy.
Pricing varies with volume, so ask for a quote based on your actual monthly hours of footage, minutes of audio and page counts.
Pick one live request from your backlog, mixed media if you have it, and run it through a trial before you commit. Log the minutes spent on each file type. Count the corrections the reviewer had to make, then compare the totals with your current process. Time saved depends on your files, so your own figure is the one to trust.
Frequently asked questions
-
It is a single platform that redacts personal data from video, audio, documents and screen recordings. Everything runs through one workflow, with one set of permissions and one audit log. A request that mixes file types can then be handled in one place instead of three.
-
No, because no model finds every face or spoken name, so a reviewer should check every export before release.
-
You can, and a team that handles only one file type probably should. The gaps appear when one request mixes formats, because each tool keeps its own settings, reviewers and records. A third-party decision then has to be made again in each tool, and the results may not match.
-
Run a real mixed request through a trial and try to recover covered text and blurred faces from the exports. Then compare the hours spent with your current process.