Home / API
Redaction API
Add redaction to your own product without building a detection stack. Upload, trigger analysis, start a job, receive status, retrieve the output.
At a Glance
REST
Versioned endpoints, OpenAPI specification
OAuth
Client credentials, plus service application credentials
Webhooks
A documented event catalogue, with delivery logs
6
Deployment postures, identical API in each
The Call Sequence
Upload
Media is posted to the API from wherever your system holds it.
Detect
Analysis is triggered against the object and PII classes you selected, with a compliance framework if the release is measured against one.
Redact
The job runs and the redaction is burned into a new rendition.
Get notified
A webhook fires against the documented event catalogue, with delivery logs, so nothing polls.
Retrieve
The redacted file is downloaded, along with its metadata, logs and custody history.
Two Shapes
Fully automated
For high-volume routine work where nothing needs a person. Everything runs in the background and your system never opens the file.
Review in the loop
For sensitive or high-risk content where someone has to sign off. The AI detects, analysts review and correct in Studio Space, then the final redaction is applied and your system retrieves the approved output.
Most agencies pick the second
A public-sector release is consequential, and the review step is what makes it defensible.
The same endpoints serve both
Choosing between them is a workflow decision, not a different integration.
What You Can Call
Media
Upload, manage and download assets
Detection
Trigger analysis, select classes and thresholds
Redaction
Start a job, apply a template, set the output policy
Retrieval
Download the redacted rendition
Metadata
Asset attributes, logs and custody history
Notifications
Webhook subscriptions against the event catalogue
Going Further Than REST
For embedding Studio Space inside your own application rather than alongside it, the platform also offers JavaScript widgets, UI extension apps, player plugins, a server object model, delegated permission passthrough and a partner app model.
Where Content Comes From
Content is pulled from the systems you already run rather than uploaded by hand, on a schedule or continuously.
Amazon S3
Buckets, including private ones
Azure Blob Storage
Containers and paths
SharePoint
Document libraries
OneDrive
Personal and shared drives
Google Drive
Folders and shared drives
Dropbox
Folders
Microsoft Teams
Meeting recordings as they are made
Zoom
Cloud recordings
NAS and SAN
File shares inside your own network
Kaltura is supported for an existing video library. Box, Webex, GoTo Meeting and Slab are available on request.
Scoping an Ingestion
A connector is scoped rather than pointed at everything.
An extension allowlist
Only the file types you name are imported, so a bucket holding a mix of content yields only what is meant to be redacted.
Excluded folders
Named paths are skipped, which is how part of a share is processed and the rest left alone.
Hierarchy or flat
The source folder structure is preserved or flattened, whichever your downstream system expects.
What happens to the source
After ingestion the original is left in place, moved, or deleted, as configured.
Content is stored on Azure, AWS S3 including private buckets, Wowza, or local storage, configured per portal, so a deployment can keep content in your own storage account. Worked examples: S3 to S3 pipelines, Microsoft 365 and the full set of integration routes.
Get the Spec
Tell us the workflow and we will send the OpenAPI spec and a sandbox.