Regulatory AlertsRegAlert for Agents
The same 54,000+ regulatory documents from ~300 financial regulators, for Claude, ChatGPT and your own agents — over MCP. Search is free. Full text costs a cent.
Add it in a minute
- Claude.ai: Settings → Connectors → Add custom connector → URL
https://mcp.regalert.today/mcp→ sign in with your RegAlert account. - Claude Code:
claude mcp add --transport http regalert https://mcp.regalert.today/mcp - ChatGPT (developer mode): add an MCP server with the same URL.
Sign-in is the normal RegAlert account (OAuth, nothing to copy). No API key to manage; the account page shows what each agent used.
Tools
| Tool | Cost | What it does |
|---|---|---|
search_documents | free | Full-text search with regulator, jurisdiction, region and topic filters. Metadata and a summary snippet per hit. |
recent_documents | free | What a regulator, jurisdiction or region published in a date range, newest first. |
list_regulators | free | The regulators covered and the codes the other tools take. |
get_document | 1 credit | The full text of one document as markdown, with metadata and summary. 50 reads a month are free; re-reading within 30 days is free. |
get_account | free | Balance, expiry, free reads left this month. |
buy_credits | free | A Stripe Checkout link for a credit pack or the bulk plan. |
Pricing
Free
$0
50 full-text reads a month. Search and metadata without limit. Sign in, no card.
Credits
1¢ per read
$5 = 500 · $20 = 2,500 · $50 = 7,500. Valid 12 months. Buy on the account page or with buy_credits.
Bulk
$199 / month
Unlimited reads, full-corpus export as JSONL, daily changes feed, licence to use the data inside your product.
Only get_document spends anything. When an account has nothing left the tool refuses and says how to top up — search keeps working.
What you get, and what you don't
Every document is captured from the regulator's official site the day it appears, with the original file preserved, the text extracted (translated to English where needed), dated, tagged by topic and summarised. The authoritative text is always the regulator's own, linked from every record.
Scraping the site is not a licensed path; the MCP server and the Bulk export are. Questions, volume beyond Bulk, or a regulator we don't cover yet: ask@regalert.today.