Browser Use connector for AI agents

Tools your web research agent reaches for on Browser Use, scoped per user.

CALL ANY TOOL

19 tools for browser automation in Browser Use: start, cancel, and follow up on agent runs, read results and event logs, manage sessions and cost caps, start and stop cloud browsers, and list downloaded and saved files.

browseruse_cancel_run

Run cancel

Cancel an in-progress Browser Use agent run so it stops working and stops incurring model charges. Returns the run's details with its status. A run that already finished is returned unchanged. Use this to stop one run. Use browseruse_stop_browser to also shut down a kept-alive browser, or browseruse_delete_session to cancel and hide the whole session.

Parameters

Name

Type

Required

Description

run_id

string

Required

ID of the agent run (a UUID), returned as id by browseruse_create_run or listed by browseruse_list_runs. Example: 6f1c2b9e-3d4a-4f8b-9c2e-7a5d1e0b8c34

browseruse_create_browser

Create browser

browseruse_create_run

Run create

Start a Browser Use cloud agent run: the agent opens a stealth cloud browser and works on a task you describe in plain language. Returns the run id, sessionId, workspaceId and status (usually queued). The run continues in the background. Use this to start new work, or pass session_id to continue an idle session with the same browser and history. Use browseruse_send_session_message to add a message to a session that may still be busy. Poll browseruse_get_run_status until status is completed, failed or cancelled, then read the answer with browseruse_get_run.

Parameters

Name

Type

Required

Description

task

string

Required

What the agent should do, in plain language. Include the website, the steps and what to report back. Example: Go to news.ycombinator.com and return the titles and URLs of the top 5 stories.

agentmail

boolean

Optional

Set true to give the agent a temporary email inbox it can send, receive and read mail with (for sign-ups or email codes). Default false.

keep_alive

integer

Optional

Seconds (120 to 14400) the session's browser stays open after the run ends so a follow-up can reuse the same page. Default 120. Kept time is billed as browser time. Example: 600

max_cost_usd

number

Optional

Budget in USD for model and web-search charges across the session. New sessions default to 25. On a follow-up, omit to keep the current budget; a value replaces it. Browser time and proxy traffic are billed separately. Example: 5

model

string

Optional

Model the agent runs on. Omit to use Browser Use's default (gpt-6-luna with high reasoning). Values change as Browser Use adds models; current examples include gpt-6-luna, gpt-6-sol, gpt-6-astra, claude-sonnet-5, claude-opus-5, gemini-3.1-pro, gemini-3-flash and the early-access presets bu-fast and bu-ultrafast. An unsupported value is rejected with a 422 error.

model_params

object

Optional

Provider-native parameters for the chosen model, forwarded unchanged, such as reasoning effort. Examples: {"reasoning": {"effort": "high"}} for OpenAI models, {"thinking": {"type": "adaptive"}} for Anthropic models, {"thinkingConfig": {"thinkingLevel": "high"}} for Google models. Pass {} to turn off the model's default parameters. Not accepted by bu-fast and bu-ultrafast.

output_schema

object

Optional

JSON Schema for a structured answer. When set, the agent returns JSON matching it in the run's output field. Pass a JSON object (or the same object as a JSON string). Example: {"type": "object", "properties": {"stories": {"type": "array", "items": {"type": "object", "properties": {"title": {"type": "string"}, "url": {"type": "string"}}}}}}

profile_id

string

Optional

Browser profile (a UUID) to load, with its saved cookies and logins. Create profiles in the Browser Use dashboard. Example: a9d2c5e8-7f1b-4e3a-8c6d-2b0f9e4a1c57

proxy_country_code

string

Optional

Two-letter lowercase ISO country code of the residential proxy the browser uses. Defaults to us. Example: de

record

boolean

Optional

Set true to record the browser to an MP4 video. The recording URL appears on browseruse_get_browser after the browser stops. Default false.

session_id

string

Optional

Continue this existing session (a UUID from a previous run's sessionId) so the agent keeps its conversation history and, while it is still open, the same browser. Omit to start a new session. Example: 2b7e4c1a-9f3d-4e6b-8a2c-5d1f0e9b7a63

workspace_id

string

Optional

Files workspace for the run (a UUID). Omit on a new session to create one; omit on a follow-up to reuse the session's workspace. Example: 8d3a6f2e-1b4c-4a9d-b7e5-0c2f9a1d6e48

browseruse_delete_session

Delete session

browseruse_get_browser

Get browser

browseruse_get_run

Run get

browseruse_get_run_events

Get run events

browseruse_get_run_status

Get run status

browseruse_get_session

Get session

browseruse_get_session_cost

Get session cost

browseruse_list_browser_downloads

List browser downloads

browseruse_list_browsers

List browsers

browseruse_list_runs

List runs

browseruse_list_sessions

List sessions

browseruse_list_workspace_files

List workspace files

browseruse_send_session_message

Send session message

browseruse_stop_browser

Stop browser

browseruse_update_session

Update session

browseruse_update_session_cost_cap

Update session cost cap

Build your Agent

Same auth pattern across LangChain, OpenAI, Anthropic, and Google ADK.

import { ScalekitClient } from "@scalekit-sdk/node";
import { createReactAgent } from "@langchain/langgraph/prebuilt";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

// Browser Use tools scoped to this user
const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["browseruse"], toolNames: [
    "browseruse_create_run",
    "browseruse_get_run_status",
    "browseruse_get_run"] },
  pageSize: 100,
});

const agent = createReactAgent({ llm, tools });
await agent.invoke({ messages: [{ role: "user", content: "List every plan and price on the competitor pricing page" }] });

import OpenAI from "openai";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const openai = new OpenAI();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["browseruse"] }, pageSize: 100,
});

const res = await openai.chat.completions.create({
  model: "gpt-5",
  messages: [{ role: "user", content: "List every plan and price on the competitor pricing page" }],
  tools,
});

// Execute the tool call with the user's vaulted Browser Use API key
await sk.tools.executeTool(res.choices[0].message.tool_calls[0], "user_123");

import Anthropic from "@anthropic-ai/sdk";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const anthropic = new Anthropic();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["browseruse"] }, pageSize: 100,
});

const msg = await anthropic.messages.create({
  model: "claude-sonnet-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "List every plan and price on the competitor pricing page" }],
  tools,
});

// Tool call runs with the user's vaulted Browser Use API key
await sk.tools.executeTool(msg.content, "user_123");

import { Agent } from "@google/adk/agents";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["browseruse"] }, pageSize: 100,
});

const agent = new Agent({
  name: "web_research_agent",
  model: "gemini-2.5-pro",
  instruction: "Browser Use tools scoped to this user",
  tools,
});

await agent.run("List every plan and price on the competitor pricing page");

Try these prompts

Copy any prompt into your agent. Each maps directly to a Browser Use tool. Click to copy, paste into your agent, done.

Agent runs

Copy the prompt

Copied

Go to our competitor's pricing page and list every plan with its price.

Copy the prompt

Copied

Check whether the run I started 5 minutes ago is done.

Copy the prompt

Copied

Show the step-by-step events of the last failed run.

Sessions and cost

Copy the prompt

Copied

Ask the pricing session to also capture the FAQ section.

Copy the prompt

Copied

What has the research session cost so far?

Copy the prompt

Copied

Set a $5 budget on the lead research session.

Browsers and files

Copy the prompt

Copied

Start a cloud browser I can control over CDP.

Copy the prompt

Copied

List the files the last browser downloaded.

Copy the prompt

Copied

Stop every active cloud browser in the project.

SEE HOW AUTH WORKS

Each user connects their own Browser Use API key once, and Scalekit sends it with every call. Keys stay vaulted, every call is scope checked, and every action is logged.

Authorize

Your user connects

Browser Use

once. We tie it to their identity and the meetings they approved — no shared bot account, no org-wide access

Who:

user ‘A’

when:

Once per user

access:

Limited to user

Store

Their

Browser Use

token lives in a vault scoped to them. User A's meetings are never reachable by an agent acting for user B, even on the same connection

vault:

encrypted

scope:

per-user

tokens:

auto-refreshed

Resolve

When your agent calls a

Browser Use

tool, we fetch the right token server-side. It never touches your agent, never appears in the LLM context, never shows up in your logs

speed:

~40ms

check:

before every call

seen by:

nobody

Audit

Every

Browser Use

tool call is logged — who triggered it, which meeting was fetched, what came back. 90 days of history, tied to the user who authorized it

history:

90 days

export:

SIEM-ready

logged:

every call

Test other agents

See the same per-user auth pattern across other AI and automation connectors.

GTM and RevOps Teams

Competitive intelligence briefing agent

Scans Gong calls for competitor mentions, matches each one to its Notion battlecard, and DMs every affected rep a single Slack digest per cycle. Every call runs as the PMM who owns the briefing, never a shared bot.

GTM and RevOps Teams

Outbound prospecting agent

Searches Apollo for prospects matching your ICP, scores and ranks them, drafts personalized outreach in Gmail, and logs every send to Google Sheets. Mail goes out as the rep, not from a shared inbox.

Why Scalekit

Secure your agent's access. Connectors ship in minutes

01.

Shared tokens break per-user analytics

A shared Browser Use API key looks fine in a demo. In production every agent run looks like one service account, and you cannot tell which user triggered it. Scalekit resolves the credential of the actual user who triggered the agent, never a shared bot.

// shared key
audit → bot_service_account

// scalekit
audit → user_abc ✓

02.

Authentication is not authorization

03.

Multi-tenancy is architectural

04.

Browser Use today. Ten connectors tomorrow.

“Our agents act across Salesforce, Gong, Google Drive, and more, on behalf of every customer. Scalekit behind the scenes meant we can keep adding tools without ever rebuilding how credentials or tool calling work.”

Venu Madhav Kattagoni

Head of Engineering / Von

FAQs

Frequently Asked Questions

Does the agent access Browser Use as the user or through a shared API key?

As the user. Each user connects their own Browser Use API key once, and Scalekit sends it with every call. Audit logs attribute every agent run and browser to that user, not a shared service account.

Where is the Browser Use API key stored?

In Scalekit's managed AES-256 token vault, namespaced per tenant. To rotate it, the user reconnects with a new API key. Revocation is a single dashboard action. Credentials never appear in prompts, logs, or LLM context.

Can I limit what the agent does in Browser Use?

Yes. Filter by tool name in listScopedTools to expose only what you want, for example run results and session costs without browseruse_delete_session, which cancels any run in progress and removes the session, or browseruse_stop_browser. Scalekit also enforces scope checks before every API call.

What happens when a user revokes Browser Use access?

The connection is invalidated on the next tool call. Subsequent requests for that user fail closed with a clear error. Other users in the tenant remain unaffected. The event is logged for audit.

Can I cap what an agent run spends?

Yes. browseruse_update_session_cost_cap sets a spending budget for a session without starting a run, and browseruse_get_session_cost reports LLM, search, browser, and proxy costs so far. browseruse_cancel_run stops a run in progress so it stops incurring model charges.