evaluation of semantic search

the easy way

launch prototype

Image Image

Why Seagle rocks

Seagle, a platform for comparative evaluation of semantic text encoding models on information retrieval (IR) tasks. Seagle offers

  • word embedding aggregators, which represent texts as algrebraic aggregations of pretrained word embeddings
  • pretrained semantic encoders
  • and allows for their comparative evaluation on arbitrary IR collections.
  • monolingual as well as cross-lingual IR collections

Watch The Video


Try our amazing evaluation test CSV sets.

Small Set 3x

3 queries for a fast tour of Seagl's full functionality

Download 3x CSV

Medium Set 20x

Explore more data insights by experiencing our queries explorer.

Download 20x CSV

Large Set 143x

143 queries and releted relevancies to reproduce evaluation results.

Download 143x CSV


Our Team


Markus Dietsche

IT-Systems Engineer Deployment


Fabian David Schmidt

Data Scientist Methods


Goran Glavas

Prof. Dr. Text Analytics


Simone Paolo Ponzetto

Prof. Dr. Information Systems