Start web or Drive research to find NEW sources. Workflow: research_start → research_poll → research_import
AI agents invoke research_start to trigger actions in Notebooklm. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool executes research operations that query external sources (web or Google Drive) and initiate data retrieval workflows. While it does not delete data (not Destructive), modify user data (not Write), or directly handle money (not Financial), it does trigger external operations and state changes in the research workflow.
From the tool's definition The tool 'research_start' initiates web or Drive research operations that 'find NEW sources' and is part of a workflow (research_start → research_poll → research_import) that performs external operations.
Documented attack patterns abuse exactly the kind of access research_start gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Notebooklm, and nothing reaches the server without passing your rules. This is the rule we recommend for research_start:
{
"version": "1",
"default": "deny",
"tools": {
"research_start": {
"limits": [
{
"counter": "research_start_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}research_start stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Start web or Drive research to find NEW sources. Workflow: research_start → research_poll → research_import. It is categorised as a Execute tool in the Notebooklm MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Notebooklm MCP server in PolicyLayer and add a rule for research_start: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Notebooklm. Nothing to install.
research_start is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the research_start rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for research_start. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
research_start is provided by the Notebooklm MCP server (moodrobotics/notebooklm-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Notebooklm, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
29 Notebooklm tools catalogued and risk-classified — across an index of 43,000+ MCP servers.