Converters

Vector Database

A real, fully client-side lexical vector console — 192-dimensional hashed-trigram vectors, cosine similarity, localStorage document store. No API calls, no AI model, no data leaves this browser.

—Collections

—Stored Documents

192Demo Dimensions

BrowserProcessing Location

What is this console?

This static build hashes character trigrams into a 192-dimensional lexical vector in your browser. It approximates textual overlap; it is not an AI-model embedding or true semantic search.

Quick Start

1) Add text under “Index Documents.” 2) Enter a comparison phrase under “Lexical Similarity.” Data stays in this browser's localStorage and no network request is sent.

Name	Description	Vectors	Model	Status

Create Local Collection

Name (slug)

Local vector profile

Description

Comparison phrase

Collection

Enter a phrase — results are ranked by trigram overlap, not meaning comprehension.

Target collection

Source tag

Documents — one local document per line

Enter two texts to compare their 192-dimensional lexical vectors with cosine similarity. The score reflects character and phrase overlap, not semantic equivalence.

Text A

Text B

Execution Location

Mode: browser-only static demo Vector: hashed character trigrams · 192 dimensions Storage: this browser's localStorage

Network Connection

External API requests: disabled External model requests: disabled Uploads: none

Important Limitation

This is a deterministic lexical demonstration. It does not run BGE-M3, MiniLM, an LLM, or any other AI model.