Get started

Overview

Find personal data in text before your app sends the text to an AI service.

The Check API reads a text and tells your server whether the text contains personal data. Your server calls it before it sends the text to an AI service. The answer has three parts.

  • result is UNSAFE when the text contains personal data, and SAFE when it does not.
  • findings lists each piece of personal data with its type and its position in the text.
  • maskedText is a copy of the text with each finding replaced by a label, such as [PHONE].

Your server decides what to send

  1. Your server

    Sends text to the Check API.

  2. Check API

    Returns result, findings, and maskedText.

  3. Your server

    Sends maskedText or the original text to the AI service.

The Check API does not block the text and does not send it to your AI service. Your server reads the answer and picks the next step. It can send maskedText, send the original text, or stop the request.

One endpoint

The Check API has one endpoint, POST /api/v1/check. Your workspace decides which types of personal data the Check API looks for. A request cannot change that list. Data types lists all 48 types.

Languages of the text

The Check API accepts text in any language. It supports Vietnamese, English, Japanese, and text that mixes these languages. A text in another language can get SAFE even when it contains personal data.

Regions

Flowmask runs in the regions in this table. Each region has its own console and its own API URL. The API of the Japan region is https://api.jp.flowmask.ai.

RegionConsole
Japanhttps://jp.flowmask.ai

A workspace belongs to one region. An API key works only in the region of its workspace.

Next steps

  • Quickstart takes you from a new account to your first 200 response.
  • Check text describes every field of the request and the response.

On this page