Teddy Search is a local semantic search tool built for professionals who need to quickly extract insights from structured documents like .xlsx, .csv, and more.
Designed for enterprise users working with corpora of ~<10k records — ideal for cybersecurity, GRC, or other specialized data analysis needs.
Powered by Sentence Transformers, a family of machine learning models that convert text into dense vector embeddings for fast and accurate similarity comparisons.
You may create a .env file at the root of the project to override default behavior:
TEDDY_SEARCH_DEFAULT_MODEL="MiniLM-L6-v2" # Must match one of the supported model keys