Skip to main content

Tools

The Tools page consolidates all tool-related settings: Built-in Tools, MCP Servers, and API/Webhook Tools (API endpoints).

Built-in Tools​

BrowseWiz gives AI models tools, which can help automate tasks. Currently supported built-in tools/toolsets are:

  • Browser Use Agent - AI agent that can navigate, click, fill forms, and extract content from web pages
  • Google Spreadsheet toolset - allows LLM to note information in a structured way
  • Google Calendar toolset - allows the LLM to fully manage a calendar: from simple event creation and rescheduling multiple events, to planning and optimizing the schedule for several weeks based on guidelines

Each tool has a switch, with a short description under its name:

SwitchDefaultNotes
Spreadsheet toolsOnNo extra marking.
Calendar toolsOffMarked Beta. Has a "Configure" button. The switch stays disabled until you set up a calendar with "Configure"; saving the setup switches the tool on (if you save again while it is on, it switches off).
Browser automation agentOnHas a "Configure" button that opens the "Configure Browser Automation" dialog. See Settings.

These are the same switches as "Enable spreadsheet", "Enable calendar" and "Enable browser agent" in the chat's Tools menu. Changing a switch in one place changes it in the other.

Browser Use Agent​

The Browser Use Agent is an AI agent that can interact with web pages on your behalf. It can read page content, click elements, fill in forms, and extract structured information.

For full details, see the dedicated Browser Use Agent documentation.

Google Spreadsheet toolset​

Allows AI to create and manage spreadsheets with specified columns. It is useful to note structured data from visited websites:

  • leads (conference participants, LinkedIn profiles, organizations)
  • products
  • job postings
  • interesting places
  • reviews
  • interesting events
  • posts or whole articles

Example usage scenarios​

Lead generation specialist​

Searching an online database of companies/organizations - set BrowseWiz context to active tab, run "add lead" prompt.

Recruiter/Talent sourcer​

Searching profiles on LinkedIn - sets BrowseWiz context to active tab and runs "add candidate profile with note…".

Market researcher​

Looking for competition products and listing all his findings in a spreadsheet with all the details and parameters in columns.

caution

The Google Spreadsheet toolset only works with spreadsheets it creates. It cannot access or manage your existing Google Sheets spreadsheets.

Google Calendar toolset​

Allows the AI to fully manage a calendar: from simple event creation and rescheduling, to planning and optimizing the schedule for several weeks based on your guidelines. Supports recurring events, attendees, notifications, and scheduling preferences.

For full details, see the dedicated Google Calendar toolset documentation.

caution

The Google Calendar toolset only works with events it creates. It cannot access or manage events from your existing Google Calendars.

MCP Servers​

To add an MCP server, fill in:

  • Server URL: the server's address, for example https://example.com/mcp.
  • Authorization header (optional): if the server needs authentication, enter the full header value, for example Bearer <token>. BrowseWiz sends it as the Authorization header with every request to the server.

Then click "Add". BrowseWiz connects to the server over Streamable HTTP. After adding the server, tools will be synchronized in BrowseWiz and ready to use by LLM models. A server with an authorization header shows an "Auth" badge. Each server has buttons to edit or delete it.

info

In case MCP server tools are updated, it is required to close and re-open (we always recommend Alt/Option+B) the BrowseWiz side panel. This way the new tool definitions will be fetched.

API/Webhook Tools​

The API/Webhook Tools section empowers your AI assistant by connecting it to external services through APIs and Webhooks. This powerful integration capability lets you extend the AI's functionality beyond conversation.

Adding a Tool​

To add a new tool, provide:

  • Tool Name: A name to identify the tool.
  • Endpoint URL: The accessible API endpoint for the tool. A request with the selected HTTP Method (POST by default) is sent here, with the tool's arguments in the request body as a JSON object matching your JSON Schema.
  • HTTP Method: The HTTP method to use (GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS). The arguments are only sent in the request body, never in the URL, so with GET or HEAD the endpoint receives no arguments. Use POST, PUT, PATCH or DELETE for a tool that needs them.
  • JSON Schema: The structure outlining how to interact with the tool: what data should be sent in request body (properties) and what should the values contain. Example:
{
  "name": "dummyName",
  "description": "Replace this placeholder description.",
  "parameters": {
    "type": "object",
    "properties": {
      "dummy_property": {
        "type": "string",
        "description": "Replace this property with your own."
      }
    }
  }
}
  • Headers (optional): Additional HTTP headers for authentication or configuration. Mark headers as "secret" to store them securely, with the input field displayed as a password field.

These tools enable your AI to perform actions like:

  • Fetching real-time data
  • Interacting with external services
  • Processing specific requests through custom APIs

Note: Make sure you have the necessary API credentials and permissions before adding a tool.

⚠️ Important: Exercise caution when adding APIs that can modify data or perform actions. The AI will make autonomous decisions about when to call these APIs based on context, and while it tries to make appropriate choices, it may not always interpret the requirements correctly. Consider using read-only endpoints or adding validation layers where possible.

Last updated on