# microsoft/BlingFire

A lightning fast Finite State machine and REgular expression manipulation library.

Repository: https://github.com/microsoft/BlingFire
Canonical: https://ross.abutalabs.com/products/blingfire
Language: C++
License: MIT
License Family: permissive
Last push: 2024-12-08T20:01:56+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2730, "days_push": 633, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1890, forks 142 (observed 2026-08-28T04:05:49.504387+00:00)

## What it is
Bling Fire is a fast finite state machine and regular expression manipulation library from Microsoft, primarily used for high-performance natural language tokenization. It supports pattern-based tokenization, WordPiece, SentencePiece Unigram LM and BPE, with precompiled models for BERT, XLNet, and other popular NLP models.

## Use cases
- tokenize text for BERT or XLNet models
- fast word and sentence segmentation of large text corpora
- run WordPiece or SentencePiece BPE tokenization in production
- stem or lemmatize words with finite state machines
- match multi-word expressions in search queries
- syllabify or hyphenate words
- tokenize text from Python, Rust, Ruby, C#, or JavaScript

## When to choose
- you need extremely fast tokenization with minimal initialization overhead
- you need consistent tokenization across multiple NLP model formats via one API
- you process large volumes of text in a low-latency or resource-constrained environment

## When to avoid
- you need to train new tokenizer models from scratch (use SentencePiece or Hugging Face Tokenizers)
- you need rich linguistic analysis beyond tokenization
- you prefer a pure high-level Python ecosystem tool with extensive community support

## Facets
- artifact type: library
- maturity: maintenance
- function: nlp, parser, machine-learning
- domain: machine-learning, developer-tools
- platform: cross-platform, python, cpp, rust, wasm
- tags: tokenizer, finite-state-machine, regular-expressions, wordpiece, sentencepiece, bpe, tokenization, sentence-breaking, stemming, natural-language-processing, nodejs

## Member repositories
- microsoft/BlingFire (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:49.504387+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:13:02.356608+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/BlingFire (fetched 2026-08-28T04:05:49.504387+00:00, sha 584854cfe800)
- Data as of 2026-08-30T08:39:29.467469+00:00.