Extract the web.
Navigate it as atoms.
Noema Extractor is a browser extension that sends any article or page to your local Noema canvas — parsed, structured, ready to explore. Available for Chrome and Firefox.
How it works
-
Open any page
Navigate to any article, blog post, or documentation page in your browser.
-
Click the icon
Hit “Extract & Send to Noema” in the extension popup.
-
Explore on the canvas
The content arrives in Noema as a structured tree of atoms, ready to navigate spatially.
What you get
Parser modes
- Syntactic Structure extracted from HTML/Markdown headings and bullets. No AI, instant.
- Semantic A lightweight AI model (Claude Haiku) decomposes unformatted content into meaningful atoms.
Everything stays local
- Content is stored in your local SQLite database.
- AI calls go directly from your desktop to the model provider.
- No cloud storage. No data collection. No intermediary.
Requirements
- Noema desktop v0.2 or later must be running on the same machine
- Available for Windows and Linux
- Download Noema →
Support
The extension says “Could not connect to Noema”
Make sure Noema desktop is open and running. The extension connects to a local server on port 11420 — if Noema is closed, the connection fails.
The extracted content looks incomplete or messy
Some pages block Readability.js extraction (heavy JS apps, paywalled content, single-page apps without server-rendered HTML). Try on a standard article page.
Which parser mode should I use?
Use Syntactic for well-structured pages (docs, Wikipedia, blogs with clear headings). Use Semantic for dense or unformatted text where you want AI decomposition.
- Switch mode in Noema’s session settings before extracting.
Does the extension work without an AI API key?
Yes — Syntactic mode requires no API key. Semantic mode requires a configured provider key in Noema settings.
For bugs or feature requests: open an issue or reach out via the contact form.