Thunderbit

Thunderbit is an AI web scraper delivered as a Chrome extension that turns pages, PDFs and images into spreadsheet tables in one click, with templates, subpage scraping, exports to Sheets, Airtable and Notion, plus an API, CLI and MCP server.

Thunderbit is an AI-driven web scraper that converts web pages, PDFs, documents and images into structured spreadsheet rows. It is aimed at sales and operations teams, along with real estate and marketing users, who need data from websites but have no interest in writing selectors or maintaining scripts. The main product is a Chrome extension, and the same engine is also exposed through an API, a command-line interface and an MCP server. In practice, a user opens a page and runs a single scrape. The AI decides how the page should be organised into columns and extracts the table without further input. When specific fields are needed, the user can describe them in plain language instead of configuring CSS selectors. The same field structure can be applied across different sites, PDFs, documents and images, which helps when sources are inconsistent. Subpage scraping lets the agent open each linked page, pull extra details and append them as new columns, replacing the habit of opening dozens of tabs and pasting into Excel. Output can be reshaped while scraping. Summaries, categorisation, translation and calculations can be added as columns, so the exported data arrives in the intended format. Long-form content such as articles and transcripts can be captured as well. Results go to Google Sheets, Airtable or Notion, or can be copied into other apps. For popular sites such as Amazon, eBay and Google Maps, pre-built templates return the same fields every time and consume no AI credits, which gives predictable output on common targets. Typical uses include contact collection, lead generation from directories and marketplaces, tracking buying signals such as hiring or funding changes, and data enrichment. The API, CLI and MCP server let agents in Claude, Cursor or other MCP-compatible hosts fetch structured page data without a browser. Among alternatives, Thunderbit sits between developer-oriented scraping frameworks and older point-and-click extractors: it trades fine-grained control for speed and low setup, which suits business users more than engineers building large crawling pipelines. A free tier is available, and the homepage cites ratings from Capterra, G2 and the Chrome Web Store.