Tabs
Open, list, focus and close tabs in the real Chrome window.
Generated from @mehmoodqureshi/chrome-mcp 0.9.21. 4 tools.
These tools work on the tabs of the Chrome you already have open, not on a separate browser window. tabs_list returns each tab with its tabId, and every other tab-scoped tool takes that id; tab_new returns the id of the tab it opens.
tabs_list works in a read-only setup. tab_new, tab_select and tab_close change what the browser shows, so they need --enable-mutations. Reading a tab's content still needs its domain on the allowlist.
Pass active: false to tab_new to open a page in the background. That is how a batch call opens many pages at once without taking over the tab you are working in, then reads them in parallel.
tabs_list
List open browser tabs.
No arguments.
{
"tool": "tabs_list",
"args": {}
}tab_new
Open a NEW tab (optionally at a URL) in the background, leaving the user's tab alone; it becomes the agent's tab, where calls without tabId go. Pass active:true to show it to the user.
| Argument | Type | Required | Description |
|---|---|---|---|
url | string | no | |
active | boolean | no |
{
"tool": "tab_new",
"args": {
"url": "<url>"
}
}tab_select
Make a tab active by tabId.
| Argument | Type | Required | Description |
|---|---|---|---|
tabId | string | yes |
{
"tool": "tab_select",
"args": {
"tabId": "<tabId>"
}
}tab_close
Close a tab by tabId.
| Argument | Type | Required | Description |
|---|---|---|---|
tabId | string | yes |
{
"tool": "tab_close",
"args": {
"tabId": "<tabId>"
}
}