anchor_navigate
Navigate to URLs
LABSCO SUMMARY
Metered by browser-hour, which changes how you write the automation.
Local browser servers fail in the sandbox because there is no Chrome; this one fails only if you run out of credit. That is the trade: reliability for a meter that runs while the browser is open. Write flows that close tabs and finish, not ones that leave a session parked. The network-requests tool is the sleeper — most browser MCPs give you a screenshot of a broken page and nothing about the failed request behind it.
What it is
Anchor runs the browsers; you send an API key in a header and drive them over MCP. There is no Chrome to install and no driver to keep alive.
What you get
- Navigation, clicking, typing and waiting for conditions
- Accessibility-tree snapshots and screenshots — the snapshot is what an agent should read, the screenshot is for you
- Tab management: open, list, switch, close
- File upload, dialog handling, PDF generation
- Network request inspection, which is how you find out why a page did not load
Requirements
An API key in the anchor-api-key header. Free gives $5 of credits a month and up to five concurrent browsers; Starter is $50/mo. Infrastructure meters separately — $0.01 per browser created, $0.05 per browser-hour, $8/GB through their proxy.
Setup effort
Paste a URL with a key — add the endpoint to your client
USE CASES
01
Filling in a flow that needs a real browser
Navigate, type, click and wait for the condition, with file upload and dialogs handled rather than blocking the run.
02
Letting the agent read the page properly
Take an accessibility-tree snapshot instead of a screenshot, so the model reads structure rather than pixels.
03
Debugging a page that will not work
Inspect the network requests behind the render before guessing at selectors.
04
Turning a page into a document
Save the rendered page as a PDF once the flow has it in the right state.
15 tools
Listed in project docs. Your client may show a different list.
anchor_click
Click elements
browser_type
Type text into inputs
browser_snapshot
Get page accessibility tree
browser_take_screenshot
Take screenshots
browser_evaluate
Execute JavaScript
browser_wait_for
Wait for conditions
browser_tab_new
Open new tabs
browser_tab_list
List open tabs
browser_tab_select
Switch between tabs
browser_tab_close
Close tabs
browser_file_upload
Upload files
browser_handle_dialog
Handle alerts/confirms
browser_network_requests
View network traffic
browser_pdf_save
Generate PDFs
BEFORE YOU INSTALL
What to know first
Tools you get
The project says 24; its docs name 15.
Requires
- an MCP client that can reach a remote endpoint
- API key (HTTP header anchor-api-key)
INSTALLATION
Connect your client
Add the endpoint to your client.
Server setup
ENDPOINT URL
https://api.anchorbrowser.io/mcpRequirements
- API key (HTTP header anchor-api-key)
