Tabductor
Get started

Getting started

Tabductor does browser work for you: signing in to portals, collecting data, downloading documents, filling in forms. You describe the task in plain language. An AI agent does it once in a real browser while you can watch, and Tabductor turns that run into a script, so every run after it takes seconds and costs a fraction of an agent run. If a site changes, the agent steps back in.

Your first workflow

  1. Add balance. Usage is prepaid. Top up under Billing; see Balance and billing.
  2. Save your sign-ins. If the task needs an account on a site, create a browser profile, open its browser, sign in, and stop it to save.
  3. Describe the task. On Workflows, write what to do, where, and what to bring back, then Create workflow. See Writing a workflow.
  4. Run it and watch. Choose your profile and Run workflow. Watch the browser live, and take control if it needs you. See Runs and live sessions.
  5. Let it repeat. Add a schedule, start it by email, or call it from the API. Results build up in the workflow's store.

Where things are

PageWhat it's for
WorkflowsYour workflows. Each has tabs: Workflow (instructions, profile, schedules), Runs, Store, Email, Share and Settings (model, webhooks, delete).
SessionsEvery browser session, live or recorded.
ProfilesSaved sign-ins and the country each browser exits in.
FilesDocuments runs saved, email attachments and your uploads.
ExploreWorkflows other people published, ready to copy.
BillingYour balance, top-ups and receipts.
SettingsDefault model, secrets, API keys and developer mode, and your account.

What it's good at

Work you repeat on websites that have no API, or no export of what you need: pulling invoices and statements, watching prices and listings, building lead lists, checking portals for changes. A task that is the same steps each time distills into a reliable script. See the use cases for examples.

Getting started · Tabductor