Solutions · By capability

Video and audio redaction with the exemption on every mark

Body-worn camera, CCTV, 911 and dispatch audio, interview recordings, voicemail, Teams meetings — reviewed, transcribed, and severed in the browser, with the statutory exemption cited on every span, and released as a burned MP4 or MP3 with a hash manifest and a time-based redaction log. All of it in your own Microsoft 365 tenant.

Start Free Trial Request a Demo

Recordings are the fastest-growing category of responsive record and the slowest to process. Four independent measurements — the Canada Border Services Agency's procurement challenge, a Washington county's 2026 fee study, a vendor's police-department customer story, and a California Supreme Court record — land in the same band: four to six analyst hours for every hour of footage, and often more. Most request-management software leaves that work outside the system entirely, in a desktop tool or a separately licensed service.

Where the request suites do offer video redaction, they integrate a separately licensed third-party tool, and by that vendor's own description the round trip returns the redacted file and an audit log. The exemption codes a reviewer applies inside the video tool are burned into the picture and listed in that tool's own report; they do not become the case file's redaction record. The time-based withholding index — what was withheld, from when to when, under which provision — lives outside the request system.

AccessPoint treats a recording as a document. It sits in the same workspace as the PDFs, opens in a viewer built for review, and every mark — a blurred face, a muted span, a distorted voice — carries the exemption, the alternates, the approval, and the audit entry that a page redaction does. The released copy burns the citation into the frame at the place in the record where the deletion is made, and the redaction log records every span.

Sound familiar?

Three moments every office that handles footage recognizes — and what each one risks.

Five hours of work per hour of footage

A body-worn camera request arrives with eleven hours of video. Someone scrubs through it in a desktop tool, blurring frame by frame, then writes down in a separate spreadsheet which exemption covered each blur — because the tool has nowhere to put it.

The citation that didn't survive the handoff

The video went out to a separate redaction service and came back as a file. The exemption codes stayed behind. When the commissioner asks why 3:17 to 4:02 was withheld, the answer lives in someone's memory.

The wrong cut, released

Two versions of the same clip sit in a folder — the internal review copy and the release copy. The unblurred one gets attached, and there is no manifest to prove which one went out.

What to look for in video and audio redaction software

Whatever you buy — including if it isn't AccessPoint — these are the capabilities that separate a blur tool from a system that can defend a release.

Time-based marks that follow the subject A mark should cover a span of the timeline and, for video, a region that tracks the person through the shot. If every second is a separate decision, the hours-per-hour ratio never comes down.
The exemption on the mark, in the case system Each span should carry its statutory authority — primary and fall-backs — inside the request record, not in a separate tool's report. That is what makes a withholding defensible on appeal.
A transcript you can redact from Time-coded, speaker-labelled transcription lets a reviewer find a name and mute the moment it is spoken, instead of listening to the whole recording twice.
Detection that proposes, never applies Face and on-screen text detection should produce proposals a person accepts or dismisses — and should run where the footage is, not in a third party's cloud.
Voice distortion, not only mute Regulators have ordered distortion rather than removal, so the words stay audible and the speaker unidentifiable. Mute and bleep alone over-redact a 911 call.
A release you can prove The burned file, a hash manifest, a per-span redaction log, and a disposition for every recording — the same evidence a page release produces.

How AccessPoint redacts video and audio

One workspace for the PDFs and the footage — these are the load-bearing pieces.

Every recording is a document

Upload up to 2 GB per file through a resumable queue, or capture a Teams meeting recording with its transcript. Recordings sit beside the PDFs on the request, assessment, incident, or complaint, with a Duration column and processing chips.

Marks that follow the subject

Draw a box, blur, or pixelate over a span; keyframe it, or let automatic tracking follow the person through the clip in both directions. Mute, bleep, or distort the voice for the sound. Withhold the whole recording where that is the decision.

Detection in your own tenant

Detect faces and on-screen text — a licence plate, a name badge, a case file on a monitor — and receive one tracked proposal per hit. Face detection runs on your own App Service, so nothing leaves the tenant, and nothing is applied until a person accepts it.

Transcript-anchored redaction

Azure AI Speech transcribes in your own subscription; Teams recordings bring their own transcript. Find a word, jump to the moment it is spoken, and let find-and-redact, your detection rules, and the AI pass propose mute marks over the spoken span.

The exemption survives to the release

Each mark cites its exemption and alternates, passes approval, and is audited like a page redaction. The released MP4 or MP3 carries the citation captioned in the frame, and the redaction log records every span, effect, and exemption.

Effort and fees that match the statute

The Fees panel estimates media review from the recording's duration — four reviewer-hours per recorded hour by default, tunable to your office — and pre-fills the fee line where your schedule bills review time.

See it in AccessPoint

Real screens from the product, running in a Microsoft 365 tenant. Click any one to enlarge.

The media viewer in AccessPoint showing a committee meeting recording with four tracked faces, each captioned with the section 14 personal privacy exemption, and the marks panel listing exemption, alternates and label Click to enlarge
The media viewer. Review a meeting recording, body-worn clip or CCTV in the browser: a tracked mark follows each face, and the exemption is captioned in the frame.
The Redactions tab listing a document’s page marks and a council meeting recording’s time-coded marks, each with the exemption cited Click to enlarge
One redaction index. Page marks and time-coded marks in one list inside the request, each with the exemption cited.
The effect chooser for a media redaction mark, offering box, blur and pixelate for the picture and mute, bleep and distort for the audio Click to enlarge
Choose the effect. Box, blur or pixelate the picture; mute, bleep or distort the voice; or withhold the whole recording.

The structural difference

The citation does not survive the handoff — except here

Elsewhere in this market, video redaction is a separate product. The request system sends the footage out and gets back a rendered file and another product's audit log; the exemption codes applied inside the video tool are burned into the picture and listed in that tool's report, not held as the case file's own redaction record. AccessPoint keeps the mark, its exemption and alternates, its approval, its audit entry, the burned caption, the release package, and the redaction log as one record that never crosses a vendor boundary. Where a statute asks that the amount withheld and the exemption be indicated at the place in the record where the deletion is made, the in-frame caption does exactly that.

One record, one tenant Footage, transcript, marks, and release copies stay in your own Microsoft 365 and Azure tenant.
A time-based withholding index The redaction log lists every span with its effect and exemption — the record no page-native tool keeps for video.
No per-seat add-on Video and audio redaction is part of the platform, not a separately licensed integration.

From upload to release

The path for a body-worn camera request:

  1. Upload the footage or capture the Teams recording. It is prepared for playback and transcribed in your own subscription.
  2. Run Detect faces and text on screen, review the proposals, and add any span, region, or voice marks by hand.
  3. Cite the exemption and alternates on each mark, route the marks through approval, and check the Redacted preview.
  4. Export the response package: the MP4 or MP3 with the marks burned in, the manifest of hashes, and the redaction log.

Go deeper

Police services & public safety Body-worn camera, CCTV, and 911 audio requests — and what regulators now expect a release to look like. Video & audio records in the docs Every control in the media viewer: marks, tracking, detection, transcripts, release, and fees. FOI request management The full request lifecycle the recording belongs to — deadlines, collection, redaction, release. AccessPoint vs. GovQA How the US public-records incumbent handles video — a per-seat third-party integration. Security & data sovereignty Where the footage lives, who can see it, and what never leaves your tenant. Product overview Every module of the platform, with screenshots.

Video & Audio Redaction Questions

What kinds of recordings can AccessPoint redact?
Video and audio in the formats your records arrive in — body-worn camera and in-car video, CCTV and dashcam exports, dispatch and 911 audio, interview-room recordings, voicemail, and Teams meeting recordings captured with their transcript. Files up to 2 GB are accepted by default through a resumable upload queue, and the released copy is rendered as MP4 or MP3 with the marks burned in.

How does redaction work on a recording?
Every mark covers a span of the timeline and, for video, a region of the picture. Choose an effect — box, blur, or pixelate for the picture; mute, bleep, or distort the voice for the sound; or withhold the whole recording — and cite the exemption and alternates exactly as you would on a page. A region can be keyframed by hand or tracked automatically through the clip in both directions. The Redacted preview plays the record as the requester will receive it, and the redaction index, exemption summary, and log all cite each span.

Can it find faces and on-screen text automatically?
Yes. Detect faces and text on screen scans the video's frames and proposes one tracked region per face and a region wherever your organization's detection patterns appear — a licence plate, a name badge, a case file on a monitor. Face detection runs entirely on your own App Service with no cloud vision service and no per-frame charge; on-screen text uses your Azure AI Document Intelligence resource. Every hit is a proposal for a person to accept or dismiss. Nothing is applied silently, and a re-scan never piles up duplicates.

Does the exemption citation go with the released file?
It is burned into the frame at the span where the deletion is made, as a caption — the exemption citation, custom text, or none, per mark. The response package also carries a redaction index and a per-request redaction log listing every span, effect, and exemption, a manifest of every file's hash, and a disposition for each recording — disclosed in full, in part, or withheld. That is the time-based withholding record most public-records tools do not keep for video.

What does it cost, and what does it need?
Video and audio redaction is part of the flat AccessPoint license — there is no per-seat video add-on. In your own Azure subscription you can add Azure AI Speech for transcripts (billed per audio hour), Azure AI Document Intelligence for on-screen text detection (billed per frame read), and a scale-to-zero media processing job for production volumes (billed per second while it runs, otherwise nothing; without it, recordings are processed on the App Service itself on a larger plan). Face detection and tracking have no usage charge. Everything runs in your subscription and region.

Flat annual pricing, published in full. Every feature in every size — access requests, privacy impact assessments, breach response, complaints, video & audio redaction, and AI Assist. No per-user or per-module fees.

  • Under 500 employees$2,990/yr
  • 500–2,000 employees$7,990/yr
  • Over 2,000 employees$14,990/yr

See full pricing Build the business case

Redact the Footage Where It Lives.

Try AccessPoint free for 30 days in your own Microsoft 365 tenant — video and audio redaction included, no per-seat add-on.

Start Free Trial Request a Demo