Rust
Embed the native library directly in your program and call the redactor through its Rust API.
Anonymize text inside your programs.
Defingerprinter makes it convenient to remove personal information and secrets from text. Run it locally, connect it to your application, and use the redacted result wherever your workflow needs it.
Source & documentation ↗Local redaction / Example
Before
Contact zqx_75@example.comAfter
Contact [EMAIL_1]An example of an email address replaced with a placeholder.
One library, many workflows
Prepare text for analysis, clean up messages before sharing them, or redact logs and diagnostic reports. Defingerprinter gives your program a reusable step for removing identifying details from text.
Structured recognizers and a local model work together to detect names, email addresses, credentials, file paths, and other sensitive details, replacing them with placeholders. Your application decides how to use the result.
Connect your language
C, C++ & Python
Use the versioned C ABI from C or C++. The Python integration recipe calls the same shared library through standard-library ctypes.
Common Lisp
Load the included ASDF system and CFFI bindings to use the redactor from Lisp.
Node.js, Go & beyond
Send text through the JSONL command-line interface. The handbook includes Node.js and Go subprocess recipes; other languages can use the same stdin/stdout protocol.
Local by design
The model runs on your CPU. Nix packages the library, model, and runtime together, with no runtime downloads or cloud processing. The implementation does not write input text to disk.
Supports English, Spanish, French, German, Italian, Portuguese, and Dutch within the API's input and output limits. Unsupported input or processing failures return no text.
Redaction is probabilistic: it reduces exposure, but cannot guarantee anonymity or catch every secret.
Source & credits
Explore Defingerprinter on GitHub ↗ for integration guides, build instructions, and the complete API contract.
Built on Rampart by National Design Studio. Defingerprinter adapts its redaction policy and uses its model and tokenizer, adding native integrations and reproducible packaging. Original Chakany Systems code is MIT-licensed; Rampart-derived material, model, and tokenizer are licensed under CC BY 4.0.