Skip to main content
Glama

Start a browser-agent task (billed)

browse
Destructive

Automate browser interactions like filling forms, using a site's search, or reaching pages behind clicks; submit a task to a hosted agent and poll the returned handle for results.

Instructions

Hand a task to a hosted browser agent that clicks and types until done: MINUTES, billed per task. For forms, a site's own search, pages behind interaction; a URL you have is extract's. After 25s it returns a handle: poll browse_status, since calling this again pays for a second task. NOTHING IS CONFIGURED FOR THIS METHOD on this machine, so every call will fail. A key in one of BROWSERBASE_API_KEY, BROWSER_USE_API_KEY, SKYVERN_API_KEY turns it on: see load_env.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "task": "<task>"
    },
    "name": "browse"
  }
}

Input Schema

NameRequiredDescriptionDefault
taskYesWhat the agent should do in the browser, in plain words.
providerNo
max_stepsNoUpper bound on the actions the agent may take.
start_urlNoThe page the agent starts on. Some backends cannot start without one, and say so.
output_schemaNoJSON Schema the answer must conform to. Omit it to get prose instead.
allowed_domainsNoKeep the agent on these domains.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.13.0

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well beyond the annotations by disclosing cost model (billed per task, MINUTES runtime), the asynchronous lifecycle (25s then a handle; re-calling pays for a second task), and a hard availability constraint (no provider key configured, every call will fail, names the exact env vars and points to `load_env`). These are the facts that determine whether an agent should call this at all, and none are derivable from readOnly/openWorld/destructive hints.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with purpose and cost, then selection rules, then lifecycle, then the blockers — a well-ordered single paragraph with no filler sentences. The usage clause ('For forms, a site's own search, pages behind interaction') is telegraphic enough to require a re-read, which costs it a point.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a billed, asynchronous, currently-unconfigured tool this covers everything an agent needs before calling: cost, duration, return shape (handle), required follow-up, and the precondition that must be satisfied via `load_env`. No output schema exists, yet the description still tells the caller what comes back.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 83%, so the schema already explains task, max_steps, start_url, output_schema and allowed_domains. The description adds no per-parameter meaning (e.g. what a backend limitation on start_url looks like, or typical max_steps values), so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a concrete verb and resource ('hand a task to a hosted browser agent that clicks and types until done') and immediately bounds scope with the contrasting case ('a URL you have is `extract`'s'). An agent can distinguish browse from extract and browse_status without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit positive triggers (forms, a site's own search, pages behind interaction), an explicit exclusion routed to a named sibling (`extract`), the follow-up tool for polling (`browse_status`), and the setup tool (`load_env`). Nothing about tool selection is left to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.