What is anonymization as a service?

Anonymization as a service is a managed pipeline you send footage to and receive it back de-identified: faces, licence plates, screens, and identifying speech automatically detected and irreversibly blurred, then confirmed by human review, with a per-file certificate of what was removed. You get the privacy outcome without building and staffing the detection, tracking, and QA yourself.

Why offer this — aren’t you a data-collection company?

It is the same pass we run on every egocentric dataset we deliver, exposed as a standalone service. We built irreversible face, plate, and screen redaction with human sign-off because our own captures happen in real homes, streets, and workplaces where bystanders end up in frame. Teams kept asking to run their own footage through it, so we made it available on its own.

What can you anonymize?

Faces and heads, licence plates, legible screens and documents, badges, distinctive marks on request, and identifying speech in audio — in video, image sets, or frame sequences. Pick the classes per project. Faces-only is the cheapest fit; the full pass covers everything that re-identifies an un-consented person.

Is the anonymization reversible?

No. Every redaction is baked into the delivered pixels — no reversible mask, no metadata toggle, no key. You can also have the un-redacted source securely destroyed, recorded in writing, so no identifiable master exists anywhere.

How do you handle our source footage and privacy?

Your choice, in writing: we return only the anonymized output and delete the source, or return both and retain nothing ourselves. Either way you receive a data-processing record suitable for a DPIA, listing what was redacted, how, and by whom.

Services · Anonymization as a service

Send us footage. Get it back with every identity removed.

The de-identification pass we run on every dataset we deliver, opened up as a service you can send your own video to. Faces, licence plates, screens, and identifying speech are detected, tracked, and irreversibly blurred — then a human confirms it, per clip, before anything comes back.

Redacts: Faces · plates · screens · speech
Redaction: Irreversible, pixel-level
Confirmed by: 100% human review
Scales with: GPU, not headcount

A first-person cycling frame with a green bounding box labelled face 0.80 drawn around the rider’s detected face. — Stage one. A vision model scans the frame and returns a bounding box around every face, with a confidence score. Nothing is altered yet — this labelled view stays internal.

A real frame at both stages. Detect returns a bounding box and a confidence score; Blur bakes the redaction into the delivered pixels.

What it is

A privacy outcome, delivered — not a tool to run yourself.

Anonymization as a service means you send footage to a managed pipeline and receive it back de-identified, with a record of what was removed. You skip building the detection, the cross-frame tracking, and the human QA that make the result trustworthy.

We built this pass because our own egocentric captures happen in real homes, streets, and workplaces, where bystanders end up in frame. The honest answer was never to pretend they do not — it was to remove them before anyone downstream sees the footage. Teams kept asking to run their own video through the same pipeline, so we made it available on its own.

It is the same discipline CloudPano has applied to spatial capture across 200+ countries for years: collect at scale, but treat the people in the scene as something to protect. The difference from a do-it-yourself model or an open-source detector is not the model — it is the irreversibility, the cross-frame tracking, the human sign-off, and the per-file certificate that lets you prove the pass ran.

The services

Pick the redaction your footage needs.

Each service targets a specific exposure. Faces-only is the cheapest fit; the full video pass covers everything that re-identifies an un-consented person.

A first-person cycling frame with the rider’s face irreversibly blurred, the rest of the scene left intact.

Face blurringBlur every face in your footage, automatically and at scale.Face blurring is the automatic detection and irreversible obscuring of human faces in images or video, so that individuals can no longer be identified. A model locates each face with a bounding box, then a blur or pixelation is baked into the output pixels rather than applied as a removable overlay.Face blurring →

A first-person video frame with a green bounding box labelled face 0.80 around the subject’s face, showing the detection stage.

Video anonymizationAnonymize a whole video, not just the obvious faces.Video anonymization is the process of removing every personal identifier from a video so no individual can be recognised — faces and heads, licence plates, legible screens and documents, and identifying speech in the audio — with the redaction baked irreversibly into the delivered file.Video anonymization →

A street scene with a parked vehicle whose registration plate is blurred while the rest of the frame is left intact.

Licence plate blurringRedact every licence plate, across every frame.Licence plate blurring is the automatic detection and irreversible obscuring of vehicle registration plates in images or video, so a vehicle and its keeper can no longer be identified. Each plate is located, tracked across frames, and blurred into the output pixels rather than masked with a removable overlay.Licence plate blurring →

Every dataset we collect already ships through this pass by default — see the automatic anonymization pipeline built into our own deliveries, and the provenance and consent chain around it.

How it works

Five stages, run on every clip.

The pass is automatic, but detection alone is never the guarantee — the human review and the per-file certificate are.

1 · Detect: A vision model runs on every frame, returning a bounding box and a confidence score for each face, head, licence plate, and legible screen or document it finds.
2 · Track: Boxes are linked across frames into stable tracks, so a subject that turns away, blurs in motion, or leaves and re-enters is covered for the entire time they are on screen — not just the frames they face the camera.
3 · Redact: Each tracked region is irreversibly blurred into the pixel data of the output. There is no separate mask layer, no metadata toggle, and no key that restores the original region.
4 · Review: A human reviewer scrubs the output and confirms no un-consented person or identifier survives in any frame. If a region is uncertain, the clip is held or dropped rather than shipped.
5 · Certify: The detector version, pass timestamp, reviewer ID, and per-file settings are recorded against the clip, so the anonymization claim is checkable per file rather than asserted in a brochure.

The guarantees

What makes it trustworthy.

Automatic blurring is common. What matters is whether it is reversible, whether it fails safely, and whether you can prove per file that it happened.

Irreversible by construction: Blur is baked into the delivered pixels. No un-redacted master is shipped alongside, and there is no reversible mask or key.
Fails closed: When a reviewer is unsure a region is fully covered, the clip is held or discarded — never downgraded and shipped to hit a deadline.
Consented subjects preserved: Anyone who signed a release — your operator, your talent — can stay in frame. The pass targets un-consented third parties, not your usable signal.
Auditable per file: Every output carries the detector version, timestamp, reviewer sign-off, and redaction settings, so a batch can be spot-checked against its own record.
You keep the source, or we destroy it: Your choice, in writing: we return only the anonymized output and securely delete the source, or return both to you and retain nothing.

FAQ

Questions teams ask first.

What is anonymization as a service?: Anonymization as a service is a managed pipeline you send footage to and receive it back de-identified: faces, licence plates, screens, and identifying speech automatically detected and irreversibly blurred, then confirmed by human review, with a per-file certificate of what was removed. You get the privacy outcome without building and staffing the detection, tracking, and QA yourself.
Why offer this — aren’t you a data-collection company?: It is the same pass we run on every egocentric dataset we deliver, exposed as a standalone service. We built irreversible face, plate, and screen redaction with human sign-off because our own captures happen in real homes, streets, and workplaces where bystanders end up in frame. Teams kept asking to run their own footage through it, so we made it available on its own.
What can you anonymize?: Faces and heads, licence plates, legible screens and documents, badges, distinctive marks on request, and identifying speech in audio — in video, image sets, or frame sequences. Pick the classes per project. Faces-only is the cheapest fit; the full pass covers everything that re-identifies an un-consented person.
Is the anonymization reversible?: No. Every redaction is baked into the delivered pixels — no reversible mask, no metadata toggle, no key. You can also have the un-redacted source securely destroyed, recorded in writing, so no identifiable master exists anywhere.
How do you handle our source footage and privacy?: Your choice, in writing: we return only the anonymized output and delete the source, or return both and retain nothing ourselves. Either way you receive a data-processing record suitable for a DPIA, listing what was redacted, how, and by whom.

Send a sample. See it come back clean.

Tell us what is in your footage and how much of it there is. We anonymize a sample, you confirm the classes and recall meet your bar, and then we run the rest to spec with a delivery date.

Request anonymization →How the pass works