MCP Tool Documentation
MCP.Scientio.Com

Tool reference

MCP.Scientio.Com exposes its capabilities to AI agents as Model Context Protocol (MCP) tools. The pages below describe every tool the server publishes, grouped by the area of the system it belongs to, so that you can understand what your agent is able to do on your behalf and what each parameter means.

Before you start

Every tool call must be authenticated with your API key, sent in the X-Api-Key request header. If you do not have a key yet you can register for a free 30 day trial key, or view an existing key. Calls made without a valid key fail with an authorization error.

Point your MCP client at the server root:

https://mcp.scientio.com

Most tools operate on a KBase (knowledge base) — a private, indexed collection of your content together with an automatically maintained knowledge graph. KBases are identified by an indexId, which you can obtain with the ListKBases tool.

Tool groups

Reading the tool pages

Each tool is listed with its exact MCP name, a description of what it does, a table of its parameters and a note on what it returns. The badges shown against a tool are the behaviour hints the server advertises to MCP clients: tools marked read‑only never change your data, and tools marked destructive create, change or remove data and should be used with care. A tool with no badge carries no hint — read its description to see whether it writes anything.

A live, always up-to-date list of the tool names and descriptions the running server advertises is shown on the home page.