ClearFactr’s MCP Server makes every model a first-class tool in the agentic AI ecosystem. AI agents discover, query, stress-test, and explain your models through natural conversation.
Create plans, add tabs, define rows, set formulas, configure time periods and scenarios. An agent can build a fully-structured financial model through conversation — complete with cross-tab references, named inputs, and time-aware columns.
List plans, tabs, rows, scenarios, and cell values. The agent can navigate the full model structure, find the data it needs, and understand how tabs connect to each other — all without the user needing to point it at a specific cell.
Retrieve any cell’s formula rendered in natural language. Instead of =B3*(1+$A$7), the agent sees “Revenue × (1 + Growth Rate).” It can explain model logic to stakeholders, audit formula chains, or answer questions about how any result is calculated.
Update input values, add new rows, adjust formulas, or restructure tabs. The agent can evolve the model as requirements change. Every modification is versioned, audited, and immediately reflected in all downstream API calls and agentic workflows.
Override any input values and compute results — without saving anything. Run multi-variate sensitivity analysis, parameter sweeps, or stress tests. The model acts as a pure, stateless function: inputs in, results back, nothing changed.
List available scenarios, switch between them, and compare results across Base, Upside, Downside, or any custom scenario — all against the same compact model. No copies. No duplicated tabs. The agent works across scenarios as naturally as it works across cells.
The model is built once, never duplicated, never annotated, and it serves as a living computational API that an agent can query, stress-test, and explain — all through conversation.
That’s a fundamentally different relationship with a spreadsheet than anything Excel offers. And it doesn’t require the model builder to do anything special to enable it. Just build your model. It’s already an API.
Your end users benefit too — they get answers faster, through natural conversation, backed by the same governed, deterministic models your risk team has already approved.
It's remarkably trivial. The ClearFactr MCP Server works with any MCP-compatible AI client — Claude Desktop, Cursor, your own custom agent, or any tool that speaks the Model Context Protocol. Connect it to your ClearFactr account, point it at a model, and start asking questions.
No SDK to install. No code to write. No model annotations to maintain.
If you have a ClearFactr model, you already have an AI-ready computation engine.
