OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Semanticfinder

Semanticfinder is a JavaScript browser application for semantic search and document chat. It uses transformers.js to calculate text embeddings and cosine similarity on the client, with WebAssembly and WebGPU support, and also offers a Chrome extension.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Semanticfinder — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a browser-based search over text can surface relevant passages even when they use different wording from the query. For very large libraries or ongoing collections, the project points users toward a vector database.

Who it is for

  • People searching personal notes or other text collections by topic rather than exact keyword.
  • Readers exploring long documents and books for relevant passages or recurring themes.
  • Developers interested in building or locally building the web app or Chrome extension.

Documented uses

  • Find passages related to a natural-language search in document text.
  • Combine semantic matching with full-text search.
  • Create and save an index for a custom text collection.
  • Use a pre-indexed example to search a large book.
Before you choose

Requirements & limits

  • Embedding generation happens before repeated searches and can take time; the README reports one to two minutes for a 660,000-character test on an older CPU.
  • WebAssembly supports the compatible models listed by the project; using WebGPU requires a WebGPU-compatible model.
  • The Chrome extension includes settings for multilingual use and minimum text-segment length.
  • The repository documents local builds using npm, including separate instructions for the extension.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install and run SemanticFinder from its repository

Start the SemanticFinder application locally from a cloned repository.

Before you start

  • A cloned SemanticFinder repository
  • npm

Steps

  1. From the repository directory, install dependencies with the command below.

    Documented command
    npm install

  2. Start the application with the command below.

    Documented command
    npm run start

  3. Optionally, create a build with the command below; the README says index.html, main.css, and bundle.js are placed in dist.

    Documented command
    npm run build

Expected outcome

The application is started with the repository's start script; the optional build produces the listed files in `dist`.

Guide limitations

  • The README does not specify browser or runtime compatibility details.
  • For larger texts or an entire library of books, the README recommends using a vector database instead.
Documentation 1 ↗
Source-based practical guide

Build and load the Chrome extension locally

Create a local extension build and load it into Chrome.

Before you start

  • A cloned SemanticFinder repository
  • npm
  • Chrome

Steps

  1. Open a terminal in the repository's extension directory.

  2. Run the command below.

    Documented command
    npm install

  3. Run the command below for a static build, or the command below for the auto-refreshing development version.

    Documented command
    npm run build
    Documented command
    npm run dev

  4. In Chrome, open chrome://extensions, select Load Unpacked, and choose the build folder.

  5. Pin the extension in Chrome to access it easily.

Expected outcome

The locally built extension is loaded in Chrome and can be accessed from its pinned toolbar icon.

Guide limitations

  • The README notes that the extension may not work for everyone and suggests opening an issue if it does not.
  • The extension options include a multilingual model choice; the default is English-only.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
do-me/SemanticFinder · Official homepage title · retained in its source language
How we review sourcesSuggest a correction