Skip to main content
This guide takes you from an API token to a redacted document.

Prerequisites

  • An Nvisy account, or a self-hosted deployment
  • An API token from API Tokens in your account settings
  • A workspace — create one in the app, or with POST /workspaces
Export your token so the examples below can pick it up: export NVISY_API_TOKEN="..."

Install an SDK

The TypeScript SDK has full API coverage and is the reference implementation. The Python and Rust SDKs currently provide configured HTTP clients.

Verify Your Connection

Workspaces, documents, pipelines, and detections are all addressed by id. A workspace’s handle is display-only — read the id from the listing above and use that in every call below.

Redact a Document

Redaction is deliberately two-phase: a detection first finds sensitive data, and nothing is modified until you redact.
1

Pick a workspace and pipeline

A new workspace starts with no pipelines. Create one with client.pipelines.createPipeline(workspaceId, ...), or skip pipelines entirely and run an ad-hoc detection against a policy.
2

Upload a document

3

Start a detection

Detections move through pending → executing → complete.
4

Wait for analysis to finish

Subscribe to the detection’s event stream. It yields the current status immediately, then each transition, and ends once the detection settles.
5

Review what was found

6

Apply redactions

Pass { edits } instead of {} to layer reviewer changes over the analysis — suppressing a false positive, retagging an entity, or adding one the analysis missed.
7

Download the result

To analyze a document without a saved pipeline, use client.detections.createAdhocDetection(workspaceId, { documentId, policyIds }) and name the policies directly.

Next Steps

Redaction Workflow

How detection and redaction fit together

API Reference

Every endpoint, generated from the spec

TypeScript SDK

Full API coverage

Audit Trails

Export detection audits for archiving