The key never leaves your browser except in the direct HTTPS request to OpenAI's own API. The tool has no backend, stores nothing, and does not send your key to Gera or any third party.

How is similarity computed?

Each text is embedded once via the OpenAI Embeddings API, then cosine similarity is computed locally between every pair of vectors. The diagonal is always 1.0 because a text is identical to itself.

What can I use this for?

Spotting near-duplicate documents before indexing, validating that distinct classes in a dataset are actually distinct, and visualizing how tightly a set of texts clusters semantically.

How many texts can I compare?

The matrix grows as N² cells, so it stays readable up to roughly 15–20 texts. All N texts are embedded in a single batched request to keep cost and latency low.

Semantic Similarity Matrix (BYO-key)

Visualize semantic similarity between texts

The Semantic Similarity Matrix embeds each of your texts using your own OpenAI API key, then renders an N×N heatmap of the cosine similarity between every pair. It turns a list of strings into an at-a-glance picture of which items are semantically close — invaluable for deduplication, dataset auditing, and cluster discovery.

How it works

All texts are sent in one batched request to the OpenAI Embeddings API. The returned vectors are kept in browser memory, and cosine similarity is computed locally for each pair: the dot product of two vectors divided by the product of their magnitudes. Values range from -1 to 1, where 1 means identical direction. Each cell is colour-coded — brighter green for higher similarity — so patterns jump out immediately. The diagonal is always 1.0.

Tips and notes

Look for off-diagonal cells that are unexpectedly bright: those are likely near-duplicates you can collapse before indexing. If two texts you expect to be unrelated score high, your embedding model may be picking up shared surface vocabulary rather than meaning. Keep the set small (under ~20 lines) so the grid stays readable. The embeddings request runs against OpenAI directly with your key; nothing is stored.