SDK v2 (Current)

Overview

Smart Search v2 is the latest generation of the Smart Search platform — rebuilt from the ground up for modularity, interoperability, and scalability. It expands the search experience beyond traditional keyword or static data searches into a system that can operate across the web, internal knowledge, and third-party applications like Google Drive, Gmail, Google Calendar, and GitHub.


Core Capabilities

Web Capabilities

Enable intelligent web retrieval through multiple specialized tools:

  • get_web_search_results — Retrieve web search results (snippets, keypoints, or summaries).

  • get_web_search_urls — Fetch only relevant result URLs.

  • get_web_page — Retrieve and process a full web page.

  • get_web_page_snippets — Extract relevant content snippets.

  • get_web_page_keypoints — Summarize key insights from a web page.

Connector Capabilities

Enable deep integration with third-party systems:

  • search_google_drive

  • search_google_mail

  • search_google_calendar

  • search_github


Key Architectural Difference — MCP Integration

The biggest improvement in v2 lies in the introduction of MCP (Model Context Protocol).

Each functionality in Smart Search v2 is implemented at the tool level, also referred to as an MCP Tool. This allows every component — such as Web Search or Connector Search — to be registered, extended, and orchestrated independently.

In practice, this means:

  • Each capability is self-contained and can be called as a discrete module.

  • Smart Search can integrate natively into agentic systems that support MCP.

  • Future capabilities can be added without modifying existing code.

Last updated