Article
FAQ
Short answers to common questions about the Load Factor MCP server.
- MCP
- Updated
Overview
Each answer is short and links to the page with the details. If your question is not here, contact us.
Is it read-only?
Yes. Every tool only reads data; nothing you or your assistant does through the MCP server can change anything in Load Factor. Clients see the tools marked read-only, so ChatGPT does not ask you to confirm each call.
Who can use it?
Load Factor users whose admin has switched on MCP access for their account. What each person sees follows their own Load Factor plan: ISOs, detail level, evidence depth and modules. See Access & permissions.
Does Load Factor see my chats?
No. The server receives only the tool calls your assistant makes: the tool name and its arguments, such as a project number or company name. It never receives the rest of your conversation, your files or the assistant's replies. Those calls are recorded, and your Load Factor admins can review them. See Privacy & request logging.
Which AI clients work?
Claude on the web, Claude Desktop, the Claude mobile apps and Claude Code; ChatGPT, as a custom MCP server, including deep research and company knowledge; and any MCP client with Streamable HTTP and OAuth sign-in. See Other MCP clients.
Can I use it on my phone?
Yes, in the Claude apps for iOS and Android: add Load Factor to your Claude account once (on the web or in Claude Desktop) and it is available there too. See Connect Claude.
Do I need to install anything?
No, for Claude and ChatGPT: they connect to Load Factor from their own servers. For Claude Code you add the server with one command. See the Quickstart.
How fresh is the data?
The MCP server reads the same data as the Load Factor web app, so answers are as current as Load Factor itself. Freshness differs by dataset: queue status reflects each ISO's latest snapshot, and nodal prices are a one-day snapshot. Ask "What data does Load Factor have, and how fresh is it?" and the assistant will check the latest dates with data_catalog.
Why are "two sources agree" counts lower than I expected?
The MCP tools count a developer there only when two or more independent strong records name it, the same badge the web app shows. Attributions that come only from the interconnecting utility, an automatic EIA owner match, an automated promotion, a single weak source or a row flagged in quality checks never count. Also check the status you asked for: the default live includes projects that are already operating, while in_queue is only the pipeline. See Writing good questions.
Can my team share one connection?
No. Each person signs in with their own Google account, gets the access of their own Load Factor account, and their requests are recorded under their name. On Claude Team and Enterprise plans an Owner can add the connector once for everyone; each member still connects individually.
A tool is missing. Why?
Tools your plan does not include are not shown to your assistant. Some serve one market only (MISO transmission plans, Virginia permits, New York distributed generation). Ask "What can I access through Load Factor?", then see Access & permissions.
What does it cost?
There is no separate sign-up: MCP access is switched on for your account by your Load Factor admin. For questions about your Load Factor plan, contact us. Your AI client needs a plan that lets you add custom connectors (Claude) or custom MCP servers (ChatGPT).