For AI & Developers

Machine-readable access to FX Pattern — hourly forex desk notes across ten major G10 pairs. Read-only endpoints, Agent Skill, and citation rules.

Quick start

FX Pattern exposes static, read-only JSON files regenerated on each hourly publish cycle. No API key is required.

Snapshot rates

/data/fx-rates.json

Ten major pair reference prices and updated timestamp.

Latest desk note

/data/latest-desk-note.json

Title, author, URL, and description of the newest hourly analysis.

Agent Skill

/SKILL.md

Instructions for AI agents on how to cite and use this site.

curl -s https://forex.doubanfx.com/data/fx-rates.json
curl -s https://forex.doubanfx.com/data/latest-desk-note.json

FX rates JSON

Endpoint: https://forex.doubanfx.com/data/fx-rates.json

Method: GET · Auth: none · Format: application/json

Returns publication-time snapshot prices for all ten covered pairs. Check the updated field before quoting levels.

Example response

{
  "schema": "https://forex.doubanfx.com/for-ai/#fx-rates",
  "site": "https://forex.doubanfx.com",
  "updated": "2026-06-05T12:00:00Z",
  "disclaimer": "Informational and educational only…",
  "source": "Yahoo Finance USD=X tickers at publish time",
  "refresh": "Hourly (UTC) with desk notes",
  "pairs": [
    {
      "symbol": "EUR/USD",
      "slug": "eur-usd",
      "key": "eurusd",
      "hub_url": "https://forex.doubanfx.com/eur-usd/",
      "price": 1.08
    }
  ]
}

Pair coverage

  • EUR/USD — Euro / US Dollar
  • GBP/USD — British Pound / US Dollar
  • USD/JPY — US Dollar / Japanese Yen
  • USD/CHF — US Dollar / Swiss Franc
  • AUD/USD — Australian Dollar / US Dollar
  • USD/CAD — US Dollar / Canadian Dollar
  • NZD/USD — New Zealand Dollar / US Dollar
  • EUR/GBP — Euro / British Pound
  • EUR/JPY — Euro / Japanese Yen
  • GBP/JPY — British Pound / Japanese Yen

Latest desk note JSON

Endpoint: https://forex.doubanfx.com/data/latest-desk-note.json

Method: GET · Auth: none · Format: application/json

Metadata for the most recent hourly desk note. Follow article.url for the full HTML article, or use RSS for history.

Example response

{
  "schema": "https://forex.doubanfx.com/for-ai/#latest-desk-note",
  "site": "https://forex.doubanfx.com",
  "updated": "2026-06-05T12:00:00Z",
  "disclaimer": "Informational and educational only…",
  "article": {
    "title": "…",
    "author": "Victoria Hale",
    "author_url": "https://forex.doubanfx.com/analysts/victoria-hale/",
    "published": "2026-06-05 12:00:00 +0000",
    "url": "https://forex.doubanfx.com/posts/2026/06/05/1200-example-slug/",
    "description": "…",
    "rss": "https://forex.doubanfx.com/feed.xml"
  }
}

Do not invent article URLs — always use values from this endpoint, RSS, or the sitemap.

Agent Skill

FX Pattern publishes an Agent Skills discovery index for automated agent onboarding.

ResourceURL
Skill document https://forex.doubanfx.com/SKILL.md
Skills index https://forex.doubanfx.com/.well-known/agent-skills/index.json
API catalog (RFC 9727) https://forex.doubanfx.com/.well-known/api-catalog

Load SKILL.md into your agent context, or register the skills index URL for discovery.

Feeds & discovery

ResourceURLUse
RSS / Atom https://forex.doubanfx.com/feed.xml Recent articles (last 30 posts)
Sitemap https://forex.doubanfx.com/sitemap.xml Full URL index
llms.txt https://forex.doubanfx.com/llms.txt Citation guide & key pages
ai.txt https://forex.doubanfx.com/ai.txt AI usage policy
robots.txt https://forex.doubanfx.com/robots.txt Crawl rules & Content-Signal

Content-Signal: search=yes, ai-input=yes, ai-train=no

Citation rules

When quoting FX Pattern in journalism, research, or AI-generated answers:

  1. Name the source: FX Pattern — https://forex.doubanfx.com/
  2. Include article title, author, publication date (UTC), and permanent URL
  3. Label prices as at publication time — not live streaming quotes
  4. State: informational only; not investment advice
Source: FX Pattern — https://forex.doubanfx.com/
Article: [title]
Author: [name]
Published (UTC): [YYYY-MM-DD HH:MM]
URL: [permanent article URL]
Note: Reference prices are snapshots at publication time. Not investment advice.

See also llms.txt and About the desk.

Limits

  • No write operations, authentication, or OAuth — read-only static files
  • No MCP server or WebMCP tools — use JSON endpoints and RSS instead
  • Rates refresh hourly; do not cache beyond the updated timestamp without re-fetching
  • Respect ai-train=no in Content-Signal for model training use