Latest, 19 Oct 2026:5 usersMove the mouse over the chart for any month
Ratings count
monthly, 8 months
Latest, 19 Oct 2026:1 ratingsMove the mouse over the chart for any month
The Chrome Web Store rounds user counts down to the leading digit, so the user line moves in steps. The ratings count is exact and is the better signal of real adoption.
About Sift
A Chrome extension that brings semantic search to any webpage. Instead of matching exact keywords like Ctrl+F, Sift understands the meaning of your query and finds the most relevant passages on the page.
How it works
To use the extension, enter your query in the text field and then press enter.
When you visit a page, Sift's content script breaks the page into text chunks and sends them to a background service worker. The service worker runs the all-MiniLM-L6-v2 sentence transformer model locally to generate embeddings for each chunk. When you search, your query is embedded and compared against the stored chunks using cosine similarity, returning the top 3 matches.
Usage tips
Sift works best when your queries are phrased as natural language questions or longer statements rather than short keywords. For example:
- "What were the main causes of the conflict?" works better than "causes"
- "How does photosynthesis produce energy?" works better than "photosynthesis energy"
This is because the underlying model is trained on sentence-level similarity, so it understands full questions better than keyword fragments.
Performance
Embedding generation runs locally in-browser, so it can take some time on large pages. Small articles process in a few seconds; very long pages (like detailed Wikipedia articles) may take longer. The model only needs to load once per session — subsequent pages will be faster.