What is the fastest way to use FXMacroData in an AI tool?

Use the hosted MCP endpoint at https://mcp.fxmacrodata.com in an MCP-compatible client, or open the FXMacroData Custom GPT in ChatGPT.

Does FXMacroData have a Codex plugin?

FXMacroData provides a Codex plugin package that wires Codex to the hosted MCP server and includes a skill for macro data workflows.

Can agents discover FXMacroData without custom setup?

Yes. FXMacroData publishes llms.txt, llms-full.txt, llms-agents.txt, ai-surface-index.json, MCP discovery, OpenAPI, and prompt-library catalogs for retrieval systems.

MCP Codex plugin LLM catalogs

Install FXMacroData where your AI agents already work.

Add official-source FX macro data to ChatGPT, Codex, Claude, Cursor, VS Code, and other MCP-compatible clients. Start with the hosted MCP endpoint, then layer prompts, LLM-readable catalogs, and the Codex plugin package when your workflow needs a richer install surface.

Connect MCP Copy prompts View AI hub

Recommended start

https://mcp.fxmacrodata.com

Use the hosted MCP endpoint for immediate agent access. Public USD evaluation works without a key for supported table workflows. Add OAuth or an FXMacroData API key for protected data, higher limits, visuals, and research tools.

Setup instructions for Claude, ChatGPT, Codex, Cursor, VS Code, and more Agent-readable manifest for autonomous retrieval systems

Install paths

Choose the right connection for the job

Fastest

Hosted MCP endpoint

Best for Claude, ChatGPT Apps, Cursor, VS Code, Codex, and any host that can add a remote MCP server.

https://mcp.fxmacrodata.com

Open MCP setup guide

ChatGPT

Custom GPT and Actions schema

Best for conversational FX macro research and OpenAI workflows that need an Actions-safe OpenAPI contract.

Open Custom GPT Open Actions schema

Codex

Codex plugin package

Best when Codex should get both MCP wiring and FXMacroData-specific skill guidance in one installable package.

codex plugin add fxmacrodata@fxmacrodata

Use this after adding an FXMacroData plugin marketplace source in Codex.

LLM-readable docs

Discovery files are part of the product surface

FXMacroData publishes plain-text, JSON, OpenAPI, and well-known discovery files so AI tools can find the right product page, prompt catalog, API contract, and MCP endpoint without guessing.

Concise LLM index https://fxmacrodata.com/llms.txt Full LLM catalog https://fxmacrodata.com/llms-full.txt Agent manifest https://fxmacrodata.com/llms-agents.txt AI surface index https://fxmacrodata.com/ai-surface-index.json MCP discovery https://fxmacrodata.com/.well-known/mcp.json Prompt catalog https://fxmacrodata.com/prompt-library.json

First workflows

Start with a useful macro task

Morning prep

Build a real-time FX event workflow

Rank the day's macro events, pull relevant release rows, and generate a structured pre-session briefing.

Risk controls

Split research from execution gating

Keep macro research separate from downstream trade controls so the data layer stays auditable.

Codex

Use Codex with FXMacroData

Let Codex inspect macro data through MCP or REST while it writes research scripts, notebooks, and checks.

FAQ

AI agent install questions

Which install path should I choose first?

Use the hosted MCP endpoint first when your AI tool supports MCP. Use the Custom GPT for ChatGPT-only research. Use the Codex plugin package when you want Codex-specific skill guidance plus MCP wiring.

What works without an API key?

Public discovery, the prompt library, LLM catalogs, and recent USD table workflows are available for evaluation. Higher-volume use, protected rows, non-USD data, COT, commodities, visual artifacts, and research tools may require OAuth or an API key.

Can I use the same setup in code and chat?

Yes. MCP is best for AI-assisted workflows, while REST and OpenAPI remain the deterministic surfaces for applications, notebooks, scheduled jobs, and generated clients.