OpenPrompt
github.com
- Free plan
- Windows
- macOS
- Linux
- Self-hosted
What OpenPrompt is
OpenPrompt is a free, open-source framework for building prompt-learning pipelines for downstream natural language processing tasks. It brings prompt templating, verbalizing, and optimization methods into a unified framework, while allowing users to develop and experiment with their own approaches. It supports pretrained language models from Hugging Face Transformers. A PromptModel combines a pretrained model with one or more templates and verbalizers for training and inference. Documented processors cover text classification, entity typing, relation classification, language inference, and conditional generation. Template options include manual, prefix, P-tuning, PTR, and mixed formats; verbalizers include one-to-one, manual, automatic, knowledgeable, PTR, generation, and soft choices. The project provides scripts to download benchmark datasets and can be installed with pip or from its GitHub source. Its listed tested requirements are Python 3.8+ and PyTorch 1.8.1+. OpenPrompt supports Linux, macOS, Windows, and self-hosted use, and is released under Apache-2.0. The README cautions that some documentation may be outdated following major project changes.
Who it suits
OpenPrompt is intended for developers, educators, and science or research users building prompt-learning methods for NLP tasks. It suits users working with the listed Python and PyTorch requirements.
Good to know, and what to watch
- Combines pretrained models, templates, and verbalizers
- Processors cover five listed NLP task areas
- Includes scripts for benchmark dataset downloads
- Apache-2.0 licensed and free
- Listed tested requirements are Python 3.8+ and PyTorch 1.8.1+
- Some documentation may be outdated after project changes
Plans
Every plan and priceOpenPrompt
Free plan
The verdict
OpenPrompt gives NLP developers a framework for assembling and experimenting with prompt-learning pipelines. Check its tested runtime requirements and the documentation caveat before choosing it.
What the maker publishes
- Core components
- A PromptModel combines a pretrained language model with one or more templates and verbalizers for training and inference. source
- Custom research
- The framework is designed to let users develop and experiment with their own prompt-learning methods. source
- Datasets
- The project provides scripts for downloading benchmark datasets. source
- Documentation caveat
- The README says some documentation may be outdated following major changes to the project. source
- Founded
- 2021 source
- Installation
- The README provides pip installation with `pip install openprompt` and instructions to install from the GitHub source. source
- Intended users
- The package metadata lists developers, education, and science or research as intended audiences. source
- License
- The repository identifies its license as Apache-2.0. source
- Maintainer organization
- THUNLP identifies itself as the Natural Language Processing Lab at Tsinghua University in Beijing. source
- Model integration
- OpenPrompt supports loading pretrained language models from Hugging Face Transformers. source
- Prompt methods
- It implements prompting methods for templating, verbalizing, and optimization in a unified framework. source
- Purpose
- OpenPrompt is an open-source framework for building prompt-learning pipelines for downstream NLP tasks. source
- Runtime requirements
- The README says the repository is tested on Python 3.8+ and PyTorch 1.8.1+. source
- Task support
- The documentation lists data processors for text classification, entity typing, relation classification, language inference, and conditional generation. source
- Template options
- The documentation lists manual, prefix, P-tuning, PTR, and mixed templates. source
- Verbalizer options
- The documentation lists one-to-one, manual, automatic, knowledgeable, PTR, generation, and soft verbalizers. source
Checked against the maker's pages on 7 Oct 2026.