ToolBunny Logo
ToolBunny
Back to Homepage

Local Embedding Similarity Calculator

Run client-side cosine similarity comparisons over text arrays safely.

Local Similarity Vector Engine:
5 items
Min Similarity Threshold20%
NxN Cosine Similarity Matrix HeatmapHover/Click cell to inspect
#S1S2S3S4S5
S1100%0%0%0%12%
S20%100%0%11%0%
S30%0%100%0%0%
S40%11%0%100%0%
S512%0%0%0%100%
Sorted Top Similar Pairs (Above 20% threshold)

About This Tool

The Local Embedding Similarity Calculator is a privacy-first vector search utility engineered for developers, data scientists, and AI researchers working with text embeddings and Retrieval-Augmented Generation (RAG) pipelines.

Comparing semantic similarity usually involves sending sensitive text data or custom embeddings to cloud services or setting up Python environments. This tool vectorizes text arrays, computes $N \times N$ cosine similarity heatmaps, ranks candidate RAG chunks against user query prompts, and evaluates raw floating-point vector arrays — 100% client-side inside your browser without network calls.

How to Use

  1. Select Engine Mode: Choose between N x N Text Matrix, Query vs RAG Candidate List, or Raw Vector Arrays.
  2. Input Data: Paste plain text sentences (one per line) or paste numerical vector arrays (e.g. [0.15, -0.42, 0.88...]).
  3. Adjust Similarity Threshold: Use the threshold slider filter to isolate top matches above your target similarity cutoff (e.g., 50%+).
  4. Inspect Heatmap Matrix: Click or hover over any cell in the similarity matrix to view cosine similarity %, dot product, and euclidean distance.
  5. Copy Code Snippets: Copy Python `scikit-learn` or `numpy` code snippets for production integration.

Common Use Cases & FAQ