Text Redaction Web Demo

Welcome, and thanks for checking out our web demo!

The purpose of this demo is to show off our system's capabilities for text identification and redaction, and we invite you to type some text into the box below to see it in action.

But before you do, a few things to note:

The data you enter in this demo are never viewed or stored by us. And in commercial deployments, your data is never sent to us in the first place.
We've added a sample API call at the bottom of this page if you'd like to dive deeper.
For a list of entities we currently redact, please see our docs.

And a few final reminders about our models:

We can provide you with a Docker image containing our models so you can easily make REST API calls to redact your data within your own environment with just a few lines of code - the data is never shared with us and it never leaves your systems.
Our models are built to run quickly (<45ms latency) on CPUs, so no need for new hardware or expensive cloud bills.
We're currently seeing 99.5%+ in-domain accuracy and support 50 different entity types, and those numbers are always improving.

Now get redacting! Simply type the text you'd like to de-identify below:

And that API call example we promised is here:

$ curl -X POST localhost:8080/deidentify_text -H 'content-type: application/json' -d '{"text": "My name is John and my friend is Grace", "key": "<Your API Key>"}'

{
"api_calls_used":1,
"output_checks_passed":true,
"pii":[
{
"marker":"NAME_1",
"text":"John",
"stt_idx":11,
"end_idx":15,
"best_label":"NAME",
"labels":
{
"NAME":0.9125
}
},
{
"marker":"NAME_2",
"text":"Grace",
"stt_idx":33,
"end_idx":38,
"best_label":"NAME",
"labels":
{
"NAME":0.8747
}
}
],
"result": "My name is [NAME_1] and my friend is [NAME_2]"
}