Sandboxed MCP server

9 actions

Connect Claude to Playwright

Navigate and automate websites in a browser sandbox. Toolspoke puts 9 of its actions behind one MCP endpoint that Claude, Cursor, and Codex all speak.

Connect View pricing

Connection: Sandboxed MCP server
Authentication: None
Actions exposed: 9
Cost per call (typical): 5 credits
Adapter: Maintained by Toolspoke

Connected in three steps

1
Install Playwright
Open the marketplace in your workspace, add Playwright to the project your agents work in, and it appears on the gateway immediately.
2
Connect the credential
Nothing to connect. Playwright needs no credential.
3
Point your agent at the gateway
Give your client one address, https://toolspoke.com/mcp. Claude Code takes it as a command, Claude and Claude Desktop add it as a custom connector, and Cursor, Codex and VS Code each read it from a config file of their own.

.mcp.json

{
  "mcpServers": {
    "toolspoke": {
      "type": "http",
      "url": "https://toolspoke.com/mcp"
    }
  }
}

One block covers every tool you have installed. Playwright shows up in the client as soon as your policy allows it, and so does everything else you install later.

Where the address goes, per client

Claude Code: Run it in your project, then /mcp to sign in
claude mcp add --transport http toolspoke https://toolspoke.com/mcp
Claude and Claude Desktop: Settings, then Connectors, then Add custom connector
https://toolspoke.com/mcp
Cursor: ~/.cursor/mcp.json, or .cursor/mcp.json for one project
{ "mcpServers": { "toolspoke": { "url": "https://toolspoke.com/mcp" } } }
Codex: ~/.codex/config.toml
[mcp_servers.toolspoke] url = "https://toolspoke.com/mcp"
VS Code: .vscode/mcp.json, or the MCP: Add Server command
{ "servers": { "toolspoke": { "type": "http", "url": "https://toolspoke.com/mcp" } } }

What Playwright asks for

Playwright needs no credential. There is nothing to go and fetch: the settings below are optional, and they are configuration rather than secrets.

User agentOptional: Optional. Overrides the User-Agent the sandbox browser sends; leave blank for Chromium's default.; Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) …

What Claude can do in Playwright

9 actions, each one declared and named by the connector rather than discovered at runtime. A workspace policy grants a person all of them, a hand-picked selection, everything on the read side, everything on the write side, or none.

Reads: 6Reads
Writes: 3Writes
Destructive: 0Destructive

Reads

Fetches data and changes nothing.

open_page
Open a URL and return the page's accessibility snapshot: headings, links, buttons and form fields with the selectors needed to interact with them. Start here before any click or form action. Pass snapshot_depth on large pages to keep the response small.
find_on_page
Open a URL and search its accessibility snapshot for text or a regular expression, returning only the matching nodes with a few lines of context. Much cheaper than open_page when you already know what you are looking for.
extract_text
Open a URL and return its rendered text content, optionally scoped to one CSS selector. Use this to read an article, a rendered price or any content that only exists after JavaScript runs.
evaluate
Open a URL and run a JavaScript function inside the page, returning its result. Use it to pull structured data out of the DOM, for example "() => [...document.querySelectorAll('h2')].map(node => node.textContent)". The code runs in the page, not on the server.
console_messages
Open a URL and return the browser console output it produced, newest last. Use it to find JavaScript errors on a page you are debugging.
network_requests
Open a URL and list the network requests the page made, with status codes. Static assets are excluded unless you ask for them, so this is a quick way to discover the JSON APIs a page calls.

Writes

Creates or updates something on the other side.

click
Open a URL, click one element and return the resulting page snapshot. The selector is a CSS selector or a ref taken from open_page. Because the browser is recreated per call, this always navigates first - chain further steps with run_flow instead of calling click twice.
fill_form
Open a URL, fill one or more fields and optionally submit, then return the resulting page snapshot. Use it for logins, searches and filter forms. Never send credentials you were not explicitly given.
run_flow
Open a URL and run an ordered sequence of interactions in one browser session, then return the final snapshot. This is the only way to chain steps, because the browser does not survive between calls - use it for multi-step flows such as accept a cookie banner, type a search, click a result.

What it will not do

Enforced by the gateway rather than left to convention, which is why each of these can be stated flatly.

It cannot call anything else: The 9 actions above are the whole of it. A call to any other name is refused before it reaches Playwright rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time.
Nothing here deletes: This connector writes to Playwright, but nothing in it deletes or permanently alters anything.
It reaches no further than your credential: Toolspoke holds no access to Playwright of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either.
It never hears from Playwright: Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Playwright. An agent has to ask.
It does not smooth over provider limits: Toolspoke does not retry, queue or back off around Playwright's own rate limits. A call that Playwright refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in Playwright?

9 named actions: 6 that only read and 3 that write. They include open_page, find_on_page and extract_text. Nothing outside that list is reachable: the connector declares each operation by name rather than proxying whatever an agent asks for.

What credentials does the Playwright connector need?

None at all. Playwright is reachable without one, so there is nothing of yours for Toolspoke to store. The connector does take user agent as optional configuration.

Does the Playwright connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 9 actions. There is a single address, https://toolspoke.com/mcp. Claude Code adds it with claude mcp add --transport http, Claude and Claude Desktop take it as a custom connector in settings, Cursor reads it from .cursor/mcp.json, Codex from ~/.codex/config.toml, and VS Code from .vscode/mcp.json. Each of them signs in to the gateway itself, so there is no key to paste.

What does the Playwright connector not do?

The 9 actions above are the whole of it. A call to any other name is refused before it reaches Playwright rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time. This connector writes to Playwright, but nothing in it deletes or permanently alters anything. Toolspoke holds no access to Playwright of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either. Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Playwright. An agent has to ask. Toolspoke does not retry, queue or back off around Playwright's own rate limits. A call that Playwright refuses comes back to the agent as a failed call.

Can I limit which actions an agent can call?

Yes, in two places. The project switches Playwright's actions on and off one at a time, for everyone in the project at once, and the screen groups them by read, write and destructive so turning off everything that deletes is one click. An individual agent key can then be narrowed further, to particular toolkits in a project and to particular actions in a toolkit. Whatever it was granted, a key never reaches a project its owner cannot.

What gets recorded when an agent calls Playwright?

Every attempt, with the agent that made it and the person that agent belongs to, the full request payload, the response payload, the status, the duration, and the credits spent. Values whose key names a secret are masked out before the record is shown to anyone. An operation the connector marks as not retained never has its response body written at all, so the gateway keeps no second copy of what was read.

More connectors

Browse the catalogue