MCP server
Access and fair use
What your assistant can read, what it costs, and how to disconnect it.
Free with an account
Every tool is free with a RegAlert account: search, lists, full texts and citation links. There is no plan to choose, no card to enter and no API key to manage.
Fair use
RegAlert is for research, not for copying the whole collection. When one account reads a very large number of different documents in full in one day, full-text reads pause until the next day (UTC). Search, lists and lineage keep working, and opening a document you already read today does not count again.
What your assistant can do
Your assistant works through the account you sign in with. It reads published documents: search results, lists, full texts and citation links.
It cannot change your RegAlert account, follow or unfollow regulators, or send alerts. It sees your own access, and never another person's account or data.
| Your assistant can | Your assistant cannot |
|---|---|
| Search and list published documents | Change your account or your alerts |
| Read the full text and the original-language text | See drafts or documents that are not published |
| Check what amended or replaced a document | See other people's accounts or activity |
| Check your own access | Spend money: there is nothing to pay |
What RegAlert records
For each call RegAlert records your account, the tool and the time — to apply fair use and to keep the service working. Your chat stays with your assistant: RegAlert sees only what your assistant sends to a tool, such as the search words. See the privacy policy.
Disconnect an assistant
Remove RegAlert in your client's connector or MCP settings. In Claude: Customize → Connectors, select RegAlert, then remove it. In ChatGPT: uninstall the RegAlert plugin. The client then stops using your sign-in. You can connect again at any time.
To close your RegAlert account, write to ask@regalert.today.
Terms
The documents are public regulatory instruments; the summaries and translations are RegAlert's. This server is the licensed way for software to read them; scraping the website is not. See the terms.