Skip to main content
Glama

run_task

Run plain-English browser tasks to scrape, fill forms, click, extract data, or log in, returning the final result, outcome, and step count from a fresh Chromium session.

Instructions

Run a browser-driven AI task on a fresh real Chromium burner. Returns the final result, the outcome and the total steps. status: completed only means the run ended: read outcome. outcome: failed with status: completed means the agent finished but could not do the task, and final_result says what stopped it. Use this when the user asks for anything web-shaped: scrape, fill, click, extract, login, place an order, search, summarize a page, etc. The browser is a real macOS-Chrome-fingerprinted Chromium with built-in CAPTCHA solving (Turnstile, hCaptcha, reCAPTCHA, ALTCHA, FunCaptcha) and behavioral humanization. Default timeout: 10 minutes.

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": "run_task"
  }
}

Input Schema

NameRequiredDescriptionDefault
llmNoLLM identifier, e.g. anthropic:claude-sonnet-5-5, openai:gpt-4o, google:gemini-2.5-flash. Omit to use the workspace's default saved key with its provider's default model; if no key is saved at all, the task runs on Browsa's hosted model and the result says so. Hosted tiers need no LLM key and add a flat credit surcharge per task (see list_models): "browsa" (default hosted model), "browsa-fast" (fastest and lightest), "browsa-smart" (for harder tasks). When you name a model, there is no hosted fallback.
taskYesPlain-English instructions for the browser agent. Be specific about what to extract and the desired output shape (JSON, list, etc.). Example: 'Go to news.ycombinator.com and return the top 3 story titles + URLs as JSON.'
countryNo2-letter egress country (e.g. US, GB, DE). Default: auto.
max_stepsNoMax agent steps. Default 25. Increase for complex multi-page tasks.
reasoningNoHow much the model reasons. Omit for the tier's or the provider's default.
llm_key_idNoId of a saved LLM key of this workspace (see list_llm_keys). Default: none. Usually not needed: with no key given, the workspace's default saved key for the provider of llm is used.
profile_idNoId of a persistent identity (agt_..., from list_identities). If set, reuses that browser identity (cookies, history, fingerprint stay sticky). Omit for a fresh burner.
initial_urlNoOptional pre-navigation URL (saves the first agent step).
ja3_profileNoTLS handshake spoof profile (e.g. macos_chrome_137). Useful for TLS-fingerprinted sites.
llm_api_keyNoPer-task LLM provider key. Rarely needed: prefer a saved key (llm_key_id, or the workspace default) so the secret never passes through the chat. Overrides llm_key_id.
timeout_minutesNoMax wall-clock minutes to wait before returning. Default 10.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.3.2

TDQS

A4.1/5.0
Behavior5/5

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

With no annotations, the description carries the full burden and does so: it discloses the environment (real macOS-Chrome-fingerprinted Chromium, CAPTCHA solving across five providers, behavioral humanization), the fresh-burner default, and the 10-minute default timeout. Most importantly it pre-empts a real failure mode by explaining that `status: completed` only means the run ended and that `outcome: failed` plus `final_result` explains what blocked the task.

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 the action, then the result semantics, then when to use it, then environment capabilities — a sensible priority order. The CAPTCHA provider list and fingerprint detail are slightly verbose but each sentence carries operational information, and nothing is redundant filler.

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

Completeness4/5

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

For an 11-parameter tool with no output schema, the description usefully explains the return shape (final result, outcome, total steps) and the status/outcome distinction, which is exactly the kind of information the schema cannot supply. It omits cost/credit implications of hosted models, which matters for a tool that spends credits per task.

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 every parameter including `llm`, `profile_id`, `ja3_profile`, and `llm_api_key` is already documented in detail. The description adds only the timeout default, which the schema already states, plus an implicit note that omitting `profile_id` yields a burner. Baseline 3 is appropriate.

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

Purpose4/5

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

The first sentence states a specific verb and resource — run a browser-driven AI task on a fresh Chromium burner — which is far more concrete than a restatement of the name. It never names a sibling such as `act`, `observe`, or `start_browser`, so an agent must infer that those are for interactive session control rather than one-shot task execution; that inference is left unstated.

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

Usage Guidelines4/5

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

"Use this when the user asks for anything web-shaped" is followed by a concrete enumeration (scrape, fill, click, extract, login, place an order, search, summarize a page), which gives a clear positive trigger. It stops short of exclusions or naming an alternative for cases where an existing browser session should be driven step-by-step instead.

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