Data extraction with BNOD

Extract structured data from any page without writing code

Point at the data you want, and BNOD pulls it into clean columns — then exports it as CSV or JSON. Tables, single fields, multi-column rows, whole catalog pages: it turns a messy web page into a structured file. It runs as a Chrome extension inside the browser you already use, so the page sees a normal logged-in person, not a bot.

Copying data by hand is slow and error-prone — you read a cell, switch tabs, paste, repeat. Cloud extraction tools promise to fix that, but they run on someone else's servers, charge per run, and get blocked when a site notices a headless browser. BNOD does the opposite: the extraction happens locally, in your session, with your real fingerprint, and the file lands on your machine.

How BNOD extracts data from a site

You show BNOD what you want and it does the rest. There are three ways to pull structured data, and you can mix them in one workflow.

Tables, straight out. The extract_table block reads any HTML <table> into clean columns, rows, and a row count — pricing tables, spec sheets, financial data, standings.
Multi-field rows. Pick a few fields on one repeating item — say a name, a price, and a rating — and BNOD extracts the same fields from every matching row on the page.
Straight from the site's API. Many pages load their data from a JSON endpoint. The http_request block hits that endpoint directly and gives you the raw structured data, skipping the rendered page entirely.
Export CSV or JSON. Every extraction lands as a downloadable CSV or JSON file — ready for a spreadsheet, a database, or another tool.
Scheduled runs. A built-in cron scheduler re-runs an extraction on any interval, so a price list or a dashboard export stays current without you opening the tab.

With 40+ blocks in the visual node editor, extraction is just one step — you can filter rows, enrich them with another request, or hand them to an AI step in the same workflow.

Why it beats cloud extraction tools

The one thing cloud tools cannot do: run inside your real browser. BNOD does. Because the extraction executes in your actual Chrome tab, it uses your real session, your real cookies, and your real fingerprint. Sites that block headless and datacenter bots see an ordinary logged-in visitor, so pages a cloud tool never loads extract normally here — including dashboards and directories that only show data once you're signed in.

It is also local-first. Your data is extracted in the browser and written to a file on your machine — it never passes through a third-party server. There is no per-run metering and no monthly platform fee. AI features are bring-your-own-key, so when a workflow uses a model you pay your provider directly at cost, not a markup.

Describe it, AI builds it

You don't have to assemble the workflow by hand. BNOD's AI Workflow Builder turns a plain-English request — "extract this pricing table and every product row on the page and save it as CSV" — into a ready-to-run workflow. It wires the blocks, adds the export step, and hands you something you can run immediately.

The builder is BYOK across 10 providers — Anthropic, OpenAI, Gemini, OpenRouter, Groq, DeepSeek, Mistral, xAI, Ollama, and any OpenAI-compatible endpoint — plus a ChatGPT Plus sign-in option. You choose the model and pay for it directly. Prefer to build by hand? Every workflow it produces is fully editable in the visual node editor.

What people extract with BNOD

Because it runs in your own logged-in browser, BNOD works as a personal data assistant for the sites you already have access to. Common jobs:

Pricing tables → CSV — pull a competitor's tiered pricing or a spec comparison into columns.
Directories → CSV — turn a filtered list of businesses, people, or listings into a structured spreadsheet.
Dashboard exports → CSV — grab the numbers off an internal report or analytics view you're signed into.
Catalog data → CSV — extract names, prices, SKUs, and ratings from a category or search page.
API JSON → CSV — read a page's underlying JSON endpoint straight into a file.

Start from a working example instead of a blank canvas — the templates gallery ships 14 ready workflows you can adapt, and you can record your own extraction with the recorder by clicking through the steps once.

How it works in 3 steps

Point. Pick a table or the fields you want on one row. BNOD detects the structure and reads every matching row.
Refine. Choose columns, add an http_request step to pull the site's raw JSON if you need it, and pick CSV or JSON — or ask the AI builder to do it from a sentence.
Run. Execute now, or set a cron schedule so the extraction repeats and the file stays fresh.

FAQ

Is it free? Yes. BNOD is free to install and extract data with. AI features are bring-your-own-key, so if a workflow calls a model you pay your provider directly — there's no BNOD subscription and no per-run fee.

Do I need to write code? No. You point at a table or the fields you want and BNOD reads every matching row — no CSS selectors and no scripts. The AI Workflow Builder can even assemble the whole extraction from a plain-English description.

Can it extract HTML tables? Yes. The extract_table block reads any HTML <table> into clean columns, rows, and a row count, ready to export.

Can it pull data from a site's API? Yes. The http_request block hits a page's JSON endpoint directly, so you can extract the raw structured data instead of scraping the rendered page.

Can it export to CSV or JSON? Yes. Every extraction can be downloaded as a CSV or JSON file, ready for a spreadsheet, database, or another tool.

Will sites block it? It's much harder to block than a cloud tool. BNOD runs inside your real Chrome with your real session and fingerprint, so a site sees a normal logged-in visitor rather than a headless or datacenter bot.

Need to pull whole lists across pages too? See web scraping with BNOD for one-click list capture with auto-pagination, and browse the templates gallery for ready-made extraction workflows. When you're ready, add BNOD to Chrome and pull your first table in minutes.