Composio Browser Tool enables AI Agents and LLMs to automate web interactions, perform web scraping, and conduct automated testing. Use cases include data extraction, form automation, website monitoring, and intelligent web navigation.
Composio Browser Tool enables AI Agents and LLMs to automate web interactions, perform web scraping, and conduct automated testing. Use cases include data extraction, form automation, website monitoring, and intelligent web navigation. On Nagent, Browser Tool is exposed as a fully-configurable ai agents integration that any agent can call — 5 actions, and no authentication authentication. No code is required to wire Browser Tool into your workflow — connect it once via the External Integrations panel and reuse it across every agent you build.
Agent builders use Browser Tool to automate the kinds of tasks ai agents teams previously handled manually. Concrete examples — each one is a single agent step in Nagent — include:
Every action and trigger is paired with a structured input/output schema (visible in the sections below), so when you wire Browser Tool into Helix — our agentic agent builder — the editor knows exactly what each step expects and produces. Configure once, deploy anywhere across your Nagent agents.
Every operation an agent can call against Browser Tool, with input parameters and output schema. Drop these into any step of an agent built in Helix.
BROWSER_TOOL_CREATE_TASKRun an AI-powered browser automation task. The AI agent will control a cloud browser to complete your task. It can: - Navigate websites, click buttons, fill forms - Extract data and return results - Handle multi-step workflows (login -> navigate -> extract) - Solve CAPTCHAs and handle dynamic content If there is no running browser session in context, call CreateTask to start one. AFTER CREATING A TASK: 1. Use GetSession with browser_session_id to get the liveUrl - share with user to watch live 2. Use WatchTask with watch_task_id to check progress and get the final result 3. Use StopTask if you need to abort TIPS FOR GOOD TASKS: - Be specific: "Click login, enter email 'test@test.com', click submit" - Break complex workflows into smaller tasks
Input parameters
Natural language instructions for the browser AI agent. Be specific and clear. Examples: 'Go to amazon.com, search for wireless headphones, and return the top 3 results with prices', 'Log into twitter.com and post a tweet saying Hello World'. For multi-step workflows, break them into clear sequential instructions.
Domain-specific credentials for automatic login. Map domain patterns to 'username:password'. Example: {'https://example.com': 'user@email.com:pass123', 'https://*.google.com': 'email@gmail.com:password'}. Wildcards (*) supported for subdomains.
URL to navigate to before starting the task. If not provided, agent starts from blank page.
Leave this empty for the first task, a new session will be auto-created. To continue in an existing browser: pass the browser_session_id from a previous CreateTask response. This preserves login state and cookies between tasks. Important: the previous task in this session must be finished or stopped (via StopTask or by waiting for WatchTask to show status='finished') before creating a new task.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
BROWSER_TOOL_GET_OUTPUT_FILEGet a download URL for a file generated by a task. When a task downloads or creates files (PDFs, images, spreadsheets, etc.), they appear in WatchTask's outputFiles list. Use this to get a download URL. The download URL is temporary (presigned) - download promptly.
Input parameters
File ID from WatchTask's outputFiles list
Task ID from CreateTask
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
BROWSER_TOOL_GET_SESSIONGet the live URL to watch a browser session in real-time. Call this right after CreateTask to get the liveUrl. Always share this URL with the user so they can watch the AI agent working in the browser. The liveUrl opens a visual stream showing exactly what the browser agent sees and does.
Input parameters
The sessionId returned from a CreateTask call you made previously. Only use session IDs from your own CreateTask responses.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
BROWSER_TOOL_STOP_TASKKill a browser task and its session. Use this when the browser agent is in a bad state: - Task is stuck or taking too long - WatchTask shows the agent going in the wrong direction - Agent is looping or making no progress After stopping, call CreateTask again to start fresh with a new browser.
Input parameters
Task ID to stop (from CreateTask)
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
BROWSER_TOOL_WATCH_TASKPoll a browser task to check progress and get results. Call this after CreateTask to monitor the task: - status='started' means still running, poll again - status='finished' means done, check output and is_success - status='stopped' means aborted - status='failed' means an error occurred Returns what the agent is doing (current_goal, current_url), screenshots from each step, and any files generated. If the agent is going wrong, use StopTask to kill it and start fresh.
Input parameters
The taskId from CreateTask response
Last step number from a previous WatchTask response (current_step value). Only steps after this will be returned for incremental updates. Omit or pass 0 to get all steps.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
No publicly available marketplace agent is found using this tool yet. There are 57 agents privately built on Nagent that already use Browser Tool.
Build on Nagent
Connect Browser Tool to any Nagent agent in minutes — no API key management, no boilerplate. Just configure and deploy.
The five questions agent builders ask before adopting a new integration.
Open the External Integrations panel inside Nagent (app.nagent.ai/externalIntegration), find Browser Tool, and click "Connect Now." You'll authenticate with no authentication (it’s public) — Nagent handles credential storage and refresh automatically. Once connected, Browser Tool is available to any agent in your workspace.
No. Nagent provides no-code integration for every tool. Once Browser Tool is connected, you configure its 5 actions directly in the agent builder UI — no API calls, no boilerplate, no schema management.
Helix — Nagent's agentic agent builder — lets you drop Browser Tool steps into any workflow visually. Pick an action (e.g., one of those listed above), fill in the inputs (Helix knows the required vs. optional schema for each parameter), and connect it to upstream/downstream steps. Triggers run as the entry point of an agent, so when a Browser Tool event fires, the agent kicks off automatically.
Every Browser Tool action and trigger ships with a fully-typed schema — input parameters with name, type, required flag, and description, plus the output payload shape. The schemas are documented in the sections above. Helix uses these schemas to validate your configuration at build time and to type-check the data flowing between steps.
Yes. While Browser Tool ships with 5 pre-built ai agents actions, you can layer custom logic around them inside Helix — pre/post-processing steps, conditional branches, retries, or stitching Browser Tool together with other connected tools. For deeper customization, talk to our team about Nagent's Agentic AI Lab — forward-deployed engineers who build Browser Tool-based workflows tailored to your business.