TabMate FAQ

TabMate is an AI browser agent and research workspace extension. It lets you ask about the page you are on with the page already in context, save excerpts and notes into a persistent workspace, and use agent mode to perform multi-step page actions automatically — clicks, fills, form interactions, and more — with built-in approval gates before anything sensitive. Research stays organized by workspace so it is easy to continue where you left off.

A chat thread is good at helping right now. TabMate is built for research you need to come back to. It keeps saved evidence, selected text, notes, prompts, and starter kits in one workspace so the next pass starts from what you already captured instead of another blank thread.

No. TabMate is designed around the page you are actively using. It works from the page you open it on or ask about. It does not passively scan background tabs or browse sites on its own.

Yes. TabMate works from the current page, so it needs page access to capture page context, selections, and supported editor content when you use it. That access supports the extension's page-context features on the pages where you actively use TabMate; it is not used for passive browsing-history collection or unrelated background scraping.

TabMate stores workspace data like notes, excerpts, saved prompts, and saved outputs so your work continues across sessions. Page context such as URL, title, selected text, and excerpts may be stored as part of that work when you choose to save or continue that research. Operational metadata including IP address and user agent is logged for authentication and security purposes.

No. TabMate does not use your data for model training.

Yes. You can delete workspace data like conversations, saved notes, prompts, and entire workspaces from within the product. You can also delete your account from the app, which disables it immediately and schedules permanent deletion of most user-owned application data. Some minimal deletion-support records may be retained longer for operational or support purposes.

Auth tokens are stored in the extension's local browser storage on your device. Server-side, we store the data necessary for your workspace to persist across sessions.

TabMate is built for recurring browser research work that needs to stay organized: page-grounded evidence, workspace notes, saved prompts, and reusable workflows you can pick up again later. If you want broad browsing and agentic help, Claude, Gemini, and ChatGPT are pushing that frontier. If you want page evidence and saved context to stay attached to the same workspace over time, that is where TabMate fits.

Yes. TabMate now includes a built-in agent mode that performs multi-step page actions automatically — clicks, fills, hovers, form interactions, and more. Action chains are bounded: the agent stops the moment it encounters ambiguity, low confidence, or an unexpected navigation. Guarded actions like form submissions always pause for your explicit approval before the agent continues. It is built around browser research but can now take action on pages you are actively working in.

Yes. TabMate applies practical guardrails around page content forwarding, stored inputs, request limits, and certain unsafe outputs. Some content may be filtered or blocked before it reaches the model or before a response reaches you. The goal is to keep your research workflow reliable without getting in the way.

TabMate fits people whose work depends on repeated browser research: product marketers, founders, students, educators, and general researchers who want page evidence, notes, and saved follow-ups to stay connected.

TabMate is available through the browser extension. Install it from the home page, sign in, and start working with page-grounded research from the current tab.