This is a service for detecting Personally Identifiable Information (PII) in text.
To run the service, download the binary on the release page and run it:
./caviarThe service will be available on port 8080.
The first time, the deep learning model will be downloaded, you will need around 600MB of free space available.
Send a POST request to /detect_pii with a JSON payload containing the text to analyze:
{
"text": "Example text containing a name like John Doe or an email address like [email protected]."
}The service will return a JSON response containing the entities detected:
{
"entities": [
{
"word": "John Doe",
"entity": "PERSON",
"score": 0.95,
"start": 30,
"end": 38,
"index": 5
},
{
"word": "[email protected]",
"entity": "EMAIL",
"score": 0.98,
"start": 54,
"end": 75,
"index": 9
}
]
}The service can be configured using environment variables. The following variables are supported:
PORT: The port to listen on (default: 8080).
This project is provided under the Commons Clause License Condition v1.0 (see LICENSE file for details) and follows the Fair-code principles. The license allows free non-production use. For commercial use or production deployments, please contact the author to discuss licensing options.
Sébastien Campion - [email protected]