Ask questions of your own sources. An agent answers them from what you give it: web pages and videos that sit in a notebook. The answer comes back in the chat.
The Agentic Part
The agent places the sources you name in a notebook, web pages, videos, texts or files, and answers every question from that notebook. To do so it connects to Gemini Notebook through an MCP server, where an MCP is an open standard that lets an AI operate tools. The agent finds which source fits a question on its own, so you do not have to search through anything. You decide which sources sit in the notebook.
Main Features
- Questions in chat, answers from your own sources
- Web pages as a source
- Videos as a source
- Texts and files as a source
- The agent places new sources in the notebook on request
Use Cases
- A team collects web pages and videos on a topic. The agent places them in a notebook and answers questions about them in the chat.
- A question comes up about a rule that sits in one of the sources. The agent answers from the sources you gave it.
What we build for you
- The sources the agent should answer from: web pages, videos, texts or files
- The access to Gemini Notebook through your Google account
- The chat in which you ask your questions
- The people who may add sources
