Do it yourself with Damaya; integrate with the ORedact API
Faces, license plates and sensitive text are detected and redacted automatically. The enterprise API is available by invitation: validate on your footage first, then agree on integration and pricing.
AI-generated scene, no real person; green marks show the faces and plates Damaya redacted
Individuals and small teams upload and process directly; teams integrating into their own systems apply for an API trial.
Self-service web app
Upload an image or video; the system detects faces, plates and sensitive text, you confirm in the editor and export.
Integration · by invitation
Bring redaction into your workflow: submit a task, poll its status, receive a callback. Validate on your footage first, then fix scope and pricing.
Scope
Available now: faces, license plates and sensitive text in video and images.Video Privacy ProcessingImage Privacy Processing
Assessed first: text and document redaction.Text Privacy ProcessingFile Privacy Processing
What Damaya and the enterprise API actually provide, plus the billing and data rules that apply on integration.
Targets: faces, license plates and sensitive text in images (phone numbers, addresses, ID numbers).
Every frame of a video is processed; moving faces and plates are covered continuously along their tracks, not only on key frames.
Review every object in the editor afterwards: keep selected people, add regions and re-render without uploading again.
Output matches the source specification with the audio track kept; 4K and HEVC sources are supported.
The API charges only for successful processing; system failures are free and automatic retries are not double-billed.
Retention and deletion of source files, caches and results are agreed in writing before go-live.
Four kinds of footage, by where it comes from.
Phone photos and videos from site visits: remove bystanders' faces and plates before sharing.
Plates and faces in insurance, repair and road records, processed in bulk inside your existing system.
Passers-by in shop visits, street footage and event videos before publishing.
Asynchronous submission, status polling and result callbacks; integrate first, then go live.
Real output frames; validate on your own samples.
An AI-generated crowd image processed by the pipeline: every face softly blurred, clothing and background unchanged. Drag the slider to compare.
The phone number on a shop sign is masked; the shop name and other text stay.
The automatic result blurs everyone; marking our own child as kept and re-rendering leaves everyone else blurred.
Output clips come from the Damaya production pipeline; the footage is phone video we shot ourselves. Some stills carry the preview watermark. The crowd, street and traffic source images are AI-generated and show no real person; the results are real production output. Text and document redaction are assessed with samples and are not shown here.
Submit a task, check its status, fetch the result, or configure a callback. Interface documentation is provided during the trial.
Give the file location and targets (faces, plates, sensitive text); optionally the masking style and a callback URL.
Poll progress or wait for the callback; queueing and status checks are free.
Download the output and object list and keep them within the agreed retention period.
# 1. Submit a task
POST /v1/tasks
{ "type": "video", "source": "<your file url>",
"targets": ["face", "plate"], "callback": "<your url>" }
# 2. Check status
GET /v1/tasks/{id}
{ "status": "processing", "progress": 62 }
# 3. Fetch the result
GET /v1/tasks/{id}/result
{ "status": "done", "output": "<download url>", "objects": 149 }Illustrative flow; endpoints and fields follow the integration documentation.
The enterprise API is available by invitation. Tell us your footage type, volume and turnaround needs; validate first, then agree on a plan.