# RegAlert for Agents

> The official publications of financial regulators worldwide — central banks, securities and insurance regulators, sanctions authorities: 80,553 documents from 373 regulators in 192 jurisdictions. Each has an English summary and the English text, on the day it is published. Free with a RegAlert account, under fair use: at bulk scale full-text reads pause until the next day (UTC); search and lineage keep working. For bulk access write to ask@regalert.today.

This file is the full reference. It is on this host because the human site (regalert.today) answers non-browser clients with a bot challenge.

## Connect

- MCP server (Streamable HTTP, stateless): https://mcp.regalert.today/mcp
- Sign-in: OAuth 2.1 with a RegAlert account; an MCP client starts it by itself from the 401 answer.
- Claude Code: `claude mcp add --transport http regalert https://mcp.regalert.today/mcp`
- Claude.ai, ChatGPT, Cursor, VS Code: add a custom connector (remote MCP server) with that URL.
- REST, JSON, the same bearer token: https://mcp.regalert.today/v1 (the routes are below; OpenAPI at https://mcp.regalert.today/openapi.json)

## How to work

1. Find documents: `search_documents` by words, or `recent_documents` by date (no input = the newest of everything).
2. Read one: `get_document uid=… query="…"` returns only the passages that match — use it for long documents. Without `query` the text comes in parts; each answer says how to get the next part.
3. See what changed it: the `lineage` block of `get_document`, or `get_lineage`, says what amended or replaced it. `in_force_status` (in_force, amended, superseded; `in_force` only means no newer amending or superseding document is known to RegAlert, not a checked commencement date; empty = no confirmed link touches the document) comes from those links. RegAlert does not check whether a document is in force: no link only means no link found.
4. Cite a document by its page: the `page` of `get_document`, or the `url` field of a list (`https://regalert.today/document/<slug>-<uid>`; a bare `/document/<uid>` redirects there).

Filters take codes, not names: `regulator` (CBN), `jurisdiction` (NG), `region`, `topic`. A wrong code is an error that names the right ones. Some regulator codes are used in several jurisdictions (FSC, FMA, CBB…): give `jurisdiction` with them.

## Tools

- `search_documents(query, regulator?, jurisdiction?, region?, topic?, licence?, in_force_only?, sort?, limit?, fields?)` — Full-text search. Each hit: uid, title, date, regulator, the first 300 characters of the summary. `total` is exact. `licence` (slug; also matches documents for all regulated firms) and `in_force_only` (drop superseded; keeps documents with no known superseding document, which does not prove they are in force) filter by what a document applies to; if the index cannot filter yet, the answer says how the hits were filtered.
- `recent_documents(regulator?, jurisdiction?, region?, topic?, licence?, in_force_only?, since?, until?, limit?, offset?, fields?)` — Newest first, with the exact total. No input = the newest documents of all regulators. Dates are YYYY-MM-DD. `licence` (slug; also matches documents for all regulated firms) and `in_force_only` (drop superseded; keeps documents with no known superseding document, which does not prove they are in force) filter by what a document applies to.
- `get_document(uid, query?, offset?, max_chars?, include_original?)` — One document: metadata, summary, lineage and the English text (markdown). With `query`: only the passages that hold those words (up to 8000 characters). Without: the text from `offset`, 40000 characters a call (most 400000). The original-language text only with include_original. A document that holds annexes or forms lists them as `enclosures`; each one is in the text under a `## <label>` heading, after the main text.
- `get_lineage(uid, direction (outgoing|incoming), relation?, limit?, offset?)` — A page of a document's confirmed citation links: what it amends, replaces, is issued under or refers to (outgoing), and what did that to it (incoming). Not a check that a document is in force.
- `list_regulators(region?, jurisdiction?, query?, all?, fields?)` — The regulators and their codes. No input = totals by region; `query` finds a regulator by part of its name or by country. `fields` adds status (ok, late, down, paused, none), last_checked_at, docs_30d, latest_published when coverage data is available.
- `list_topics()` — The topic tags the `topic` filter takes, with the number of documents of each.
- `get_account()` — This account's access: whether full-text reads are available right now.

Answers are TOON (https://toonformat.dev), a compact text form of JSON: `name[n]{a,b}:` opens a table of n rows with the columns a and b; `next[n]:` lists the calls that follow from the answer. An error is `error`, `code`, and often `suggestions`.

## Fields

- Document rows carry uid, title, published, regulator, summary by default (`summary` = the first 300 characters; `summary_full` = all of it).
- `fields` of search_documents: uid, title, published, regulator, summary, summary_full, identifier, regulator_name, jurisdiction, jurisdiction_name, region, topics, url, in_force_status, applies_scope, applies_licences, effective_date.
- `fields` of recent_documents: those and instrument_no, type, language, source, file_url, thumbnail_url, effective_kind, ended_on, ended_by, applies_activities.
- `fields` of list_regulators: code, name, jurisdiction, jurisdiction_name, region, documents, website, status, last_checked_at, docs_30d, latest_published. status, last_checked_at, docs_30d (documents published in the last 30 days) and latest_published come from the coverage data and are left out when it is not available.
- What a document applies to and when: `effective_date` and `effective_kind` (date, on_issue, relative, staged, none_found), `in_force_status` (`in_force` = no newer amending or superseding document is known to RegAlert; not a checked commencement date), `ended_on` / `ended_by` (when superseded), `applies_scope` (all_regulated, classes, named_entity, general_public, none), `applies_licences` and `applies_activities` (slugs). All are empty until a document has been checked. Search hits carry in_force_status, applies_scope, applies_licences and effective_date once the index has them.
- Filters `licence` (a slug; also matches documents for all regulated firms) and `in_force_only` (true = drop superseded; keeps documents with no known superseding document, which does not prove they are in force) work in search_documents and recent_documents. If the search index cannot filter by them yet, search filters the returned hits (or applies no filter) and the answer carries a `note` saying so.

## Codes

- Regions: Africa, APAC, CentralAsia, EU, LATAM, MENAP, NA, SEA, SouthAsia, USA.
- Topics: securities, disclosure, licensing, capital, credit, governance, payments, aml, consumer, insurance, fx, enforcement, monetary, kyc, resilience, infosec, sanctions, crypto, fraud, esg, islamic-finance, inclusion, privacy, safeguarding, market-abuse, ai, open-banking, stablecoins.
- Jurisdictions: ISO country codes (NG, SG, GB, US; a US state is US-NY). Regulators: `list_regulators query="<name or country>"`.

## REST

Send `Authorization: Bearer <token>`. Answers are JSON; an error is `{ "error": "<code>", "message": "…" }` with `suggestions` when a code was wrong. The inputs and defaults are those of the tools.

- `GET /v1/search?q=…&regulator=&jurisdiction=&region=&topic=&licence=&in_force_only=true&sort=relevance|newest&limit=&fields=`
- `GET /v1/recent?regulator=&jurisdiction=&region=&topic=&licence=&in_force_only=true&since=&until=&limit=&offset=&fields=`
- `GET /v1/documents/{uid}?query=&offset=&max_chars=&include_original=true`
- `GET /v1/documents/{uid}/lineage?direction=outgoing|incoming&relation=&limit=&offset=`
- `GET /v1/regulators?region=&jurisdiction=&query=&all=true&fields=`
- `GET /v1/topics`
- `GET /v1/account`

## 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. Research aid, not legal advice.
Docs for people: https://mcp.regalert.today/docs/overview · One-click installs: https://regalert.today/agents · Privacy: https://regalert.today/privacy · Contact: ask@regalert.today