Integration · Developer tools

Connect Claude & ChatGPT to Anchor Browser

Ask your AI to read a page that needs JavaScript, collect data from a website, or take a screenshot of a page. It runs web tasks in plain language in Anchor's cloud browser, starts and controls browser sessions, and saves pages as PDFs. You can also check usage and answer the agent when it needs a human.

Connect Anchor Browser free

Free plan, no credit card. Takes about a minute.

Try asking

“Get the text of this pricing page. It only loads with JavaScript.”
The AI opens the URL in a cloud browser and returns the page content.
“Go to this shop page and collect every product name and price.”
It runs a web task from your prompt and returns the data in the shape you asked for.
“Take a full-page screenshot of our homepage.”
It captures the URL at full height, with an optional width and wait time.

What your AI can do

Web tasks in plain language

Run a web task from a prompt, create reusable tasks with input and output schemas, and check their status and results.

Browser sessions

Start sessions with saved profiles, open URLs, click, type, take screenshots and list tabs, downloads and recordings.

Pages, PDFs and usage

Fetch page content, take screenshots, save pages as PDFs, and read billing usage and session history.

“Save this article as a landscape PDF.”
It renders the URL as a PDF, with paper size, margins and landscape if you want them.
“Start a session with my saved profile and open our admin page.”
It starts a cloud browser session with the saved profile, gives you the live view link and opens the URL.
“Is any session waiting for me?”
It lists sessions where the agent needs a human, for example for a login step, and passes on your answer so it can continue.

26 tools for Anchor Browser

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

click_in_sessionChanges dataClick in a browser session by `session_id`, either on a CSS `selector` or at `x`/`y` coordinates.
create_reusable_taskChanges dataCreate a reusable browser task from a plain-language `task_prompt` with a `task_name`, optional `input_schema` and `output_schema`. Returns a task ID to run with run_reusable_task.
end_sessionDeletesEnd a running browser session by `session_id` and stop its billing.
fetch_webpageChanges dataOpen a `url` in a cloud browser and return the page content. Works for pages that need JavaScript.
get_billing_usageGet billing usage between `from_date` and `to_date`, grouped by `granularity` (hour, day, week, month).
get_sessionGet one browser session by `session_id` with its status, settings and usage.
get_session_historyGet session usage history over time between `from_date` and `to_date`, grouped by `granularity`.
get_session_start_statusCheck a session start request by `request_id` and get the new session ID and live view link once ready.
get_task_generation_statusCheck whether a reusable task (`task_id`) has finished being generated and is ready to run.
get_task_run_statusGet the status and result of a reusable task run by `run_id`.
get_web_task_statusCheck an asynchronous web task by `workflow_id` and get its result once finished.
list_active_sessionsList the currently running browser sessions, filterable by `tags`, `profile_name`, `domains` or creation time.
list_pending_interventionsList sessions where the AI agent is waiting for a human (for example to solve a login step).
list_profilesList saved browser profiles (stored logins and cookies) that sessions can reuse.
list_session_downloadsList files downloaded during a browser session by `session_id`.
list_session_pagesList the open tabs (pages) of a browser session by `session_id`.
list_session_recordingsList video recordings of a browser session by `session_id`.
open_url_in_sessionChanges dataNavigate an open browser session (`session_id`) to `url`.
respond_to_interventionChanges dataAnswer an agent's request for human help in session `session_id` with `response` (and the `request_id` if given) so it can continue.
run_reusable_taskChanges dataRun a reusable task by `task_id` with its `input_params`; `sync` true waits for the result, otherwise check with get_task_run_status.
run_web_taskChanges dataHave an AI agent do a task on the web in plain language (`prompt`), optionally starting at `url` and returning data shaped by `output_schema`. With `async` true it returns an ID to check with get_web_task_status.
save_page_as_pdfChanges dataRender a `url` as a PDF, with optional paper size, margins, `landscape` and header/footer.
screenshot_sessionTake a screenshot of what a running browser session (`session_id`) currently shows.
start_browser_sessionChanges dataStart a cloud browser session, optionally with an `initial_url`, timeout, recording, proxy or a saved profile in `browser`/`session`. Returns a request ID to check with get_session_start_status.
take_screenshotChanges dataTake a screenshot of a `url`, with optional `width`, `height`, `wait` time and full-page capture (`capture_full_height`).
type_in_sessionChanges dataType `text` into the focused field of a browser session by `session_id`.

Set up in three steps

1
Pick the app
Create a free PipMCP account and choose Anchor Browser from the app list.
2
Paste your key
Paste your Anchor API key. Sign in to Anchor at app.anchorbrowser.io, open the API key page and copy your API key.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in Anchor Browser?

It can run web tasks in plain language, fetch page content, take screenshots, save pages as PDFs, and create and run reusable tasks. It can start, control and end browser sessions, answer requests for human help, and read sessions, profiles, downloads, recordings, billing usage and session history.

Does the AI see my Anchor API key?

No. Your Anchor API key is encrypted at rest and never shown to the AI. After you save it, it is not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on. Starting sessions, running tasks and clicking or typing in a session can require your approval, and so can ending a session. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT go to Settings › Apps & Connectors › Advanced and turn on Developer mode, then add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

PipMCP has a free plan with no credit card required. Paid plans bill per completed task. You also need your own Anchor account. Browser usage is billed by Anchor, and the AI can show you that usage.

What happens when the browser agent gets stuck?

Some steps, like a login, need a person. The AI can list sessions where the agent is waiting for a human and pass on your answer so the task can continue. You can also take a screenshot of the session to see what it shows.