Query Matcher

Semantic search powered by Potion embeddings

Model Corpus Search Query Number of Results (top-k)

Performance Testing

Select Indexes to Test

Loading indexes...

Iteration Count

Concurrent Requests

Warmup

Run each index for 5 seconds before testing

Search Terms (one per line)

Evaluation

Evaluates search accuracy using BEIR SciFact dataset (5,183 docs, 1,109 queries with relevance judgments). Builds indexes for each model and measures Recall, MRR, NDCG, and Precision.

Existing Indexes

Loading indexes...

Build New Index

Model Corpus File (.txt)