explosion/spacy-llm

What is it?

Key features

• 01Serializable llm component that adds prompt-based steps to a spaCy pipeline.
• 02Modular task and model functions, including custom ones registered via spaCy's registry.
• 03Supports OpenAI, Cohere, Anthropic, Google PaLM and Azure AI APIs.
• 04Runs open-source models from Hugging Face such as Llama 2, Mistral and Falcon.
• 05Built-in tasks for NER, text classification, relation extraction, summarization and more.
• 06Map-reduce handling for inputs that exceed the LLM's context window.

Use cases

• Prototyping an NER or text classification step before collecting labelled training data.
• Combining LLM-based components with rule-based or supervised spaCy components in one pipeline.
• Extracting structured fields from unstructured text using a hosted or local open-source LLM.

Quick start

1. 1

   Install the package with pip

   python -m pip install spacy-llm
2. 2

   Create an OpenAI API key and set it as an environment variable

3. 3

   Run text classification with the llm_textcat factory in Python

Things to know

• !Still experimental, so interface changes may break in minor version updates.
• !Most built-in models need an API key from a paid provider such as OpenAI, or a hosted LLM.

Topics

FAQ

What is spacy-llm?

A spaCy extension that plugs large language models into structured NLP pipelines for tasks like NER and text classification without training data. spacy-llm integrates Large Language Models into spaCy, letting you add LLM-powered components to a regular spaCy pipeline. Each component combines a task (the prompt template and the parser that turns the model's free-text reply into structured output) with a model backend, so results land in standard spaCy fields such as doc.ents or doc.cats.

Is spacy-llm free to use? What license does it have?

spacy-llm is open source under the MIT license. Check the license terms before using it in commercial products.

Is spacy-llm still maintained?

Actively maintained. The last commit to spacy-llm was on Sep 27, 2026.

What can I use spacy-llm for?

Prototyping an NER or text classification step before collecting labelled training data. Combining LLM-based components with rule-based or supervised spaCy components in one pipeline. Extracting structured fields from unstructured text using a hosted or local open-source LLM.

How do I install spacy-llm?

Install the package with pip: python -m pip install spacy-llm Create an OpenAI API key and set it as an environment variable Run text classification with the llm_textcat factory in Python

Where can I get spacy-llm?

The source code and installation instructions are on GitHub: https://github.com/explosion/spacy-llm