bytesense

Detect the encoding. Validate every byte. Keep your data intact.

bytesense detects text encodings with zero runtime dependencies and optional Rust acceleration. Final detection results validate the selected encoding against all supplied bytes. File and stream APIs bound their input spool by moving larger inputs to a temporary file; decoder buffers and model data are additional.

from bytesense import from_bytes

raw = "café 世界".encode()
result = from_bytes(raw)
assert result.encoding == "utf_8"
assert result.complete and result.bytes_validated == len(raw)

Start with the quick start, explore the API, or review measured accuracy and performance.

Encoding detection is inference. Full decode validation rules out an invalid codec; it cannot disambiguate every legacy encoding. Confidence is an evidence score, not a calibrated probability.