Introducing the AllRatesToday MCP Server for AI, LLMs and Agents

Real-time currency data for AI agents

Large language models are confident, fluent — and badly out of date. Ask Claude or ChatGPT for today's USD/EUR rate and you'll usually get a number from whenever the model was trained, with no warning that it's stale.

The AllRatesToday MCP Server fixes that. It plugs into any AI application that speaks the Model Context Protocol — Claude Desktop, Claude Code, ChatGPT, Cursor, Windsurf, VS Code's MCP client, and any custom agent — and gives the assistant live access to the AllRatesToday currency API. The model stops guessing and starts quoting real numbers.

What is MCP?

The Model Context Protocol is an open standard from Anthropic that lets AI applications connect to external data sources and tools through a single, consistent interface. Instead of bolting a custom integration onto every chat client, you publish an MCP server once and every compatible host — Claude, ChatGPT, Cursor, Windsurf, VS Code — can use it.

For data providers, that means one server, every assistant. For end users, it means tools just appear inside the chat they already use.

Capabilities

@allratestoday/mcp-server exposes four tools backed by the AllRatesToday API:

• get_exchange_rate — current mid-market rate between any two of 160+ supported currencies.
• get_historical_rates — time series over 1 day, 7 days, 30 days or 1 year.
• get_rates_authenticated — multi-target rates in a single call (e.g. EUR against USD, GBP and CHF at once) with higher quotas.
• list_currencies — the full set of supported ISO 4217 currency codes and metadata.

All four tools require an AllRatesToday API key, set as ALLRATES_API_KEY in the server config. The free tier covers 300 requests per month and takes about 30 seconds to register for.

Key use cases

• Research and analysis — let the model pull a year of USD/JPY and reason over the volatility instead of hallucinating it.
• Travel and treasury — "How much is 1,250 GBP in SGD right now? And in JPY?" — answered with live mid-market rates, not last year's snapshot.
• Engineering and prototyping — generate working currency-conversion code inside the same chat that just verified the rate against the production endpoint.
• Automation and agents — chain rate lookups into longer agent workflows (invoicing, pricing, FX exposure monitoring) without writing a custom HTTP client.
• Education — explore real markets through natural-language conversation rather than a static textbook table.

Getting started

1. Register a free API key at allratestoday.com/register.

2. Add the server to your AI client. In Claude Code, that's one command:

   claude mcp add allratestoday -- npx -y @allratestoday/mcp-server
   claude mcp env allratestoday ALLRATES_API_KEY=art_live_xxxxx

   For Cursor, Claude Desktop, ChatGPT Desktop and Windsurf, drop the same npx command into the client's mcp.json config — full per-client setup is in the README.

3. Ask your assistant a live currency question. It will pick up the tool, call AllRatesToday, and answer with a real number.

The server is open source on GitHub, published as @allratestoday/mcp-server on npm, and listed in the official MCP registry. A Python sibling, allratestoday-deepseek, wires the same data into DeepSeek's function calling.