Ask Claude About Your Trading (MCP)

TTMT exposes a read-only connector that lets Claude — or any MCP-compatible AI tool — answer questions about your own trading data. You point Claude at TTMT once, sign in with your TTMT account, and from then on you can ask it things like "how did my Aggressive profile do last month?" or "which of my accounts are halted right now?" in plain language. Claude fetches the numbers from your account and answers.

It can read everything the dashboard shows you — trades, positions, performance, settings, channels, risk status — and it can even preview a hypothetical order plan. What it can never do is place a trade, change a setting, close a position, or move a single dollar. The connector has no way to act on your account. It only reads.

Availability

The connector requires API access enabled on your account — the same gate as the Developer API. It is rolling out gradually. If sign-in or the consent step fails, your account may still be getting enabled. Message support and we'll get you set up.

What is MCP?

MCP (the Model Context Protocol) is a standard way for AI apps to plug into an outside source of data. Instead of you copying numbers into a chat by hand, the AI app connects directly to the source and pulls what it needs to answer your question. TTMT publishes an MCP connector for your trading data, so a supported AI tool can read it on demand — with your permission, and read-only.

Because MCP is an open standard, this works with more than one app. You can use it from:

• claude.ai (the web app)
• Claude Desktop
• Claude Code (the terminal tool)
• Any other MCP-compatible client that supports the modern connector transport and sign-in

You connect the same TTMT connector URL in whichever tool you prefer, sign in once, and ask away.

What you can ask

Once connected, you talk to Claude in normal language. It works out which of your data it needs, fetches it, and answers. A few examples:

• "Give me my morning trading summary."
• "How does my Aggressive profile compare to Balanced over the last month?"
• "Walk me through what happened on trade trd_abc123."
• "Why didn't my last EURUSD signal get traded?"
• "What's my total floating P&L right now?"
• "If I got a XAUUSD BUY signal right now, what would my order plan look like?"

That last one is a preview. Claude asks TTMT to compute the exact order plan the platform would produce for that hypothetical signal and your current settings. No trade is placed — it's a calculation, nothing more.

What's inside the connector

The connector gives Claude a fixed set of read-only capabilities. Nothing here can change your account.

Capability	Count	What it means
Read-only tools	19	Individual questions Claude can answer — list your accounts, look up a trade, summarise performance, resolve your settings, check risk halts and exposure, review drawdown and session timing, preview an order plan, and more.
Resources	5	Reference material Claude can pull in: a concepts glossary, the full settings-field reference, your live account list, and direct links to a single trade or channel.
Guided prompts	7	One-click routines — a morning review, a risk review, a weekly report, a settings healthcheck, a drawdown postmortem, and channel deep-dive / vetting.
Rate limit	60 / min	Up to 60 requests per minute, shared across all your connected AI clients.
Daily limit	Essential ~25 / day	On the Essential plan the connector allows about 25 tool calls (roughly three questions) per day, reset daily. Pro and higher plans are unlimited. See Security & Limits.

Read-only guarantee

Claude can read your data. It can never place, modify, or close a trade, change a setting, or move funds — and it can never see another user's data. There is no write capability in the connector at all. Every request is signed in as you and re-checked on the server.

To see the full list of the 19 tools and exactly what each one returns, see Tools & Capabilities. For how sign-in, permissions, and data isolation work, see Security & Access.

How access works

You never create a separate password or key for the connector. It uses your normal TTMT sign-in:

1. You add the connector URL in your AI client.
2. Your client opens a TTMT sign-in page.
3. You sign in with your TTMT account.
4. You approve a consent screen granting read-only access.
5. You're connected.

Every request after that is re-verified against your sign-in — there are no long-lived sessions to leak. You can disconnect at any time from your AI client's own connector settings.

If your account hasn't had API access enabled yet, the sign-in or consent step will fail even with a valid TTMT login. That's the same gate the Developer API uses — message support to get it turned on.

Get connected

Ready to try it? The Quickstart walks you from zero to your first answer — adding the connector in claude.ai, Claude Desktop, or Claude Code, signing in, and asking your first question.

Start here → MCP Quickstart

Related pages

• MCP Quickstart — connect your AI client and ask your first question.
• Tools & Capabilities — the full catalog of read-only tools, resources, and prompts.
• Security & Access — how sign-in, permissions, and data isolation work.
• Developer API — the read-only REST + SSE API for scripts and bots (same access gate).