Annotation of biomedical entity names
the best open-source solution
Open Source
  • Use, change and distribute
  • Social development
High Performance*
  • BioCreative: 87,54%
  • JNLPBA: 73,05%
High-end Techniques
  • Linguistic dependency parsing
  • Model combination
Flexible and Scalable
  • Extensible architecture
  • Fast annotation
Easy to Use
  • Automated scripts
  • Java library
  • Creative Commons License
  • Non-commercial use
* Overall F-measure results achieved using the evaluation methods of the respective challenges.


Gimli is a machine learning-based solution for biomedical Named Entity Recognition (NER), which goal is to automatically extract names of biomedical entities from scientific text documents. Currently, Gimli supports the recognition of gene/protein, DNA, RNA, cell line and cell type names.
In summary, Gimli receives raw text as input, and provides text with specific annotations as output.
  • Machine Learning: Conditional Random Fields (CRFs);
  • Features: orthographic, morphological, linguistic parsing and conjunctions;
  • Combination: combination of models with different orders and parsing directions;
  • Post-processing: parentheses correction and abbreviation resolution.


Get the latest official release of Gimli.
Source code:
Get a copy of the project using the following git command:
git clone git://


Full documentation:
Complete information about alternative downloads, installation and usage.
API Javadoc:
Detailed classes, methods and propreties description.

Join us

We have several ideas to make Gimli the most complete and efficient tool for biomedical information extraction. You are welcome to join us and contribute to the development of new and improved features. Please contact us: