Open the app No code. No selectors. No babysitting.
Draw the route from a web page to the fields you want. Press run. Get a table you can search, sort and download, kept in your own database.
Dashed nodes are plugins, coming soon
How it works
Say where to start, how far to follow links, and which links to follow. Patterns keep the crawl on the pages you want.
Every page becomes clean text. Lists, tables and detail pages survive intact.
Describe the record and list its fields. A model returns one tidy record per page, and says null instead of guessing.
Records land in your database and in a table you can search, sort and download.
Settings in plain language
Pages to read, how deep to follow, which links to keep, and a field list instead of a JSON schema. Everything the engine supports is still one click away under More options.
*/catalog**/product?id=*Data you can read
Search, sort, page, open any row for every field, then download CSV or JSON. Lists show as chips, empty values fade, and right-to-left text just works. Try it: this table is live.
| name | category | price | stock | rating | tags |
|---|
Made to look good
Drag to compare. Every node shows its state while it runs.
Bring your own AI agent
Connect Claude Code or Codex. Even a workflow with branches, joins and stored data: open it and nodes arrive one by one while the agent studies the site, validates, runs and fixes.
A token acts as you. Keep it secret and revoke it when done. Connect your agent →
PuzzleFox Coming soon
PuzzleFox brings authorized CAPTCHA handling and human-verification workflows into the fetcher, so a pipeline can pass the gate and keep going. It drops in as one more node, right after the fetch.
For authorized use only: where you have permission or a site’s terms allow it.
Plugins Coming soon
42 plugins in the works, each one a ready-made source that plugs straight into your pipeline.
Dark-web discovery, monitoring and threat intelligence for security teams, with continuous alerts, history and search.
Reliable access to sites behind anti-bot protection through authorized methods: detection, rendering, sessions, backoff and analytics.
A thousand exits, one intelligent network. Routes traffic across approved proxy endpoints and regions, with health monitoring, cost tracking and strict access policy.
Authorized CAPTCHA handling and human-verification workflows
Prices, market data, financial reports, indicators
Tokens, blockchain data, exchanges, transactions
Products, prices, reviews, sellers, availability
Public profiles, posts, reels, captions, engagement
Posts, threads, hashtags, accounts, trends
Videos, channels, subtitles, comments, metadata
Professional profiles, companies, jobs, public posts
Research papers, authors, citations
Repositories, contributors, issues, releases
Communities, discussions, comments
Product catalogs, variants, pricing
Speech transcription, speaker identification, audio metadata
Videos, creators, captions, trends, engagement
Public channels, groups, messages, media
Authorized chats, business messages, attachments
Teams, executives, organizational structures
Articles, headlines, publishers, topics, events
Exchange rates, historical FX data, market movements
Products, categories, prices, discounts
Products, specifications, prices, reviews
Price comparison, sellers, product matching
Authorized candidate discovery and talent intelligence
Job listings, employers, salary information
Startup profiles, funding, investors
Startups, hiring, founder information
Repositories, projects, merge requests
Preprints, abstracts, research categories
Citation graphs, scholarly metadata
Patents, inventors, assignees
Questions, answers, tags
Product launches, makers, categories
Apps, reviews, rankings, metadata
Monitors latency, reliability, cost and endpoint health across your proxy network.
Places, geographic data, POIs
Historical pages, archived content
Events, speakers, organizations
Videos, streams, creators
API references, guides, changelogs
Plugins are in development. Use them on public data or with authorization, and follow each platform’s terms.
Nine building blocks
Reads one page, many, or everything within a depth.
urlFetcher Turns noisy pages into clean text.
markdownConverter Extracts a record from each page, or answers a question.
llmProcessor CSS selectors and JSON paths. No model, no cost.
dataTransformer Keeps records and images in your own database.
dataSource Ask what you stored, with a filter or in plain words.
query Search, sort, page and download your records.
table Bar, line, pie, doughnut and scatter.
chart Show an answer or any text on the canvas.
text