Skip to main content
Glama

Run a goal autonomously

browser_run

Automates multi-step browser tasks by handing a concrete goal to a decision model that plans, executes, and verifies each action, returning status, trace, cost, and final page text.

Instructions

Hand over a full goal: Jev decides each step, code executes and verifies it, up to max_steps. Returns status, a per-step trace (operation, target, confidence, goal/stuck probabilities), usage.est_cost_usd, and the final page text. Statuses: done | goal_achieved | needs_text | needs_value | stuck | blocked | max_steps | timeout | error. Follow-ups: needs_text/needs_value → fill the reported ref with browser_type/browser_select, then browser_run again with the same goal (page state persists). stuck/blocked → browser_observe/browser_extract and decide yourself. Supply values for any fields the goal needs. Best default for multi-step tasks; use browser_act when you want to review each step.

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": {
      "goal": "<goal>"
    },
    "name": "browser_run"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlNoOpen this URL first; otherwise the current page is used.
goalYesWhat to accomplish, in one or two sentences. Be concrete, e.g. "Find the price of the Pro plan and stop when it is visible."
valuesNoField label → text to type, for any fields the task needs. Contents never go to Jev.
max_charsNoCap on returned final page text (default 6000).
max_stepsNoMax Jev decisions (default 20); each costs a fraction of a cent.
max_secondsNoWall-clock budget in seconds (default 120).

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.4/5.0
Behavior4/5

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

With no annotations, the description carries the full burden and does substantial work: it enumerates every terminal status, discloses the returned payload (per-step trace with confidence and goal/stuck probabilities, usage.est_cost_usd, final page text), states that page state persists across follow-up calls, and notes per-step cost. It does not address the safety profile of autonomous action — whether it may submit forms, make purchases, or perform irreversible operations — which is a real gap for an unannotated autonomous agent.

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?

Dense but front-loaded: the core mechanic comes first, then return shape, then status handling, then the sibling comparison. The status/status-follow-up sentences are terse fragments but each carries actionable routing information; little is wasted, though the single-block layout makes the status-to-action mapping harder to scan than a short list would.

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?

This is a complex autonomous tool with six parameters, a nested object, and no output schema, yet the description supplies the return shape, the full status vocabulary, persistence semantics, and follow-up handling for every non-terminal status. An agent has enough to invoke it once and to react correctly to each outcome.

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 100%, so the schema already explains url, goal, values, max_chars, max_steps and max_seconds. The description only reinforces `values` ('Supply `values` for any fields the goal needs') and does not add syntax, precedence, or interaction semantics beyond the schema — 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?

The description states a specific action (hand over a full goal, agent decides each step, code executes and verifies, bounded by max_steps) rather than restating the name. It also contrasts itself with the sibling browser_act, so an agent can pick between autonomous and step-by-step modes without opening either 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?

It gives an explicit default recommendation ('Best default for multi-step tasks'), an explicit alternative with the condition that selects it ('use browser_act when you want to review each step'), and conditional routing for each failure status: needs_text/needs_value → browser_type/browser_select then re-run, stuck/blocked → browser_observe/browser_extract. Nothing 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.