# fastino-ai/GLiNER2

Unified Schema-Based Information Extraction

Repository: https://github.com/fastino-ai/GLiNER2
Canonical: https://ross.abutalabs.com/products/gliner2
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-24T15:40:11+00:00

## Health v2 (maintenance only)
Score: 85/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 99, longevity 30
- inputs: {"age_days": 422, "days_push": 9, "days_rel": 9, "gap_med": 12, "n_releases_24m": 12}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1813, forks 169 (observed 2026-08-28T04:05:39.393294+00:00)

## What it is
GLiNER2 is a Python library for schema-based information extraction and text classification, supporting named entity recognition, structured data extraction, relation extraction, and span attributes in a single local model. It offers two extraction architectures (span and boundary) behind a unified AutoExtractor API and runs efficiently on CPU without external dependencies.

## Use cases
- extract named entities from text
- classify text with custom labels
- extract structured records from documents
- detect PII in text locally
- extract entity relations into typed graphs
- run information extraction on CPU without GPU
- build training data for extraction models

## When to choose
- you need schema-driven NER and classification in one model
- privacy requires fully local inference with no external calls
- you want fast CPU-based extraction without GPU infrastructure
- you need entities, relations, and classification in a single forward pass

## When to avoid
- you need generative LLM-style free-form extraction beyond schema constraints
- you require state-of-the-art accuracy on complex domain-specific extraction that fine-tuned large models provide
- your stack is not Python 3.10+

## Facets
- artifact type: library
- maturity: active
- function: nlp, machine-learning, parser, llm-inference
- domain: machine-learning, data-science, privacy
- platform: python, cli, cross-platform
- tags: named-entity-recognition, information-extraction, text-classification, relation-extraction, schema-driven, cpu-inference, local-inference, pii-detection, transformers, natural-language-processing

## Member repositories
- fastino-ai/GLiNER2 (main) score 85

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:39.393294+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 2026-08-30T03:21:10.017975+00:00, confidence not recorded.
  - readme: https://github.com/fastino-ai/GLiNER2 (fetched 2026-08-28T04:05:39.393294+00:00, sha b9e1b02f819e)
  - registry_pypi: https://pypi.org/pypi/gliner2/json (fetched 2026-08-29T10:59:55.350703+00:00, sha 676401563db7)
- Data as of 2026-08-30T08:39:29.467469+00:00.