Local Vector Embeddings Tester (ONNX)

Embedding Tester computes cosine similarity between text pairs using neural embeddings running entirely in your browser. Embeddings are dense vector representations of text — semantically similar phrases produce vectors that point in similar directions, so their cosine similarity is high. This tool uses the all-MiniLM-L6-v2 model via ONNX Runtime Web to generate embeddings locally without sending data to any server.

The all-MiniLM-L6-v2 model converts each text input into a 384-dimensional dense vector. Cosine similarity is computed between the two vectors: 1.0 means identical direction (very similar), 0 means orthogonal (unrelated), and -1 means opposite directions. The model runs in a Web Worker via ONNX Runtime Web, keeping the UI responsive during inference.