Skip to content
AIRO

Claude + MCP

Route optimization from Claude

AIRO connects to Claude through MCP, the Model Context Protocol. Claude becomes the conversational interface; the AIRO engine does the deterministic route optimization.

What you can ask Claude to do

  • Validate an uploaded delivery file and summarize the issues
  • Adjust the column mapping
  • Optimize routes for a given number of vehicles
  • Check job status and retrieve results
  • Compare two scenarios and report the differences
  • Create an Excel or CSV export for drivers

How the connection works

You authorize Claude to access your AIRO workspace using OAuth. Claude then has access to a defined set of tools, each scoped to your company. Every request is authenticated, rate limited, logged and, for anything that creates or changes data, idempotent.

Claude does not receive your full spreadsheet. It receives summaries, validation results, KPIs and links returned by the tools.

Claude is the interface, not the optimizer

Language models are not designed to guarantee optimal or even valid routes. In this design, Claude decides which tool to call and explains the result in plain language. The route computation, validation and constraint handling run in a dedicated solver that produces the same output for the same input.

Not tied to the chat

Everything is also available in the web portal, so you can review routes on the map and download exports without an assistant at all.

Frequently asked questions

Does AIRO work with Claude?

Yes. AIRO is designed to connect to Claude through MCP (Model Context Protocol). Once connected, you can ask Claude to validate an import, launch an optimization, compare scenarios or export driver sheets. Claude calls the AIRO engine, which performs the optimization.

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants call external tools securely. In AIRO, MCP exposes actions like "validate import" and "optimize routes" to Claude or ChatGPT. The assistant is the interface; the AIRO backend does the work.

Is my customer address data used to train AI models?

No. Customer files are processed by the AIRO optimization engine, not used as training data. The full contents of your uploaded file are never sent to an LLM. When you use an AI assistant, it only receives summaries and results returned by the platform's tools.

Can I optimize routes from an Excel file?

Yes. AIRO is built around the spreadsheet you already use. Upload an Excel (.xlsx) or CSV file with your orders and vehicles, map your columns, fix any validation issues, and run the optimization. You do not need to reformat your file into a rigid template.

Join the waitlist

Get early access and launch updates. Double opt-in, unsubscribe anytime.