FICHA · MANJARO

python-tiktoken

A fast BPE tokeniser for use with OpenAI's models

  • Library
  • Dev
  • Dependency only
official+codex · reviewed · May 28, 2026 description in en

Description

Tokenizes text with fast byte-pair encoding routines used to estimate or prepare input for OpenAI model families. It helps developers count tokens, split text, and understand model context usage before sending requests.

Use it as a library in AI tooling. Token counts are not a privacy filter, so review content, secrets, and user data separately before sending text to external services.

Permissions

Permissions not analysed for this source yet.