MONEI for AI agents
How an AI agent works with MONEI: through the MONEI MCP server, or by reading this documentation directly.
MONEI MCP server
Start here. The MONEI MCP server connects an AI agent to a MONEI account and to this documentation:
https://mcp.monei.com
- The user signs in to MONEI (OAuth) and chooses one account and Read or Write access.
- The agent reads and manages that account: payments, refunds, customers, subscriptions and more.
- The agent also searches this documentation, so it needs no other server.
https://mcp.monei.com/testconnects to the test account: simulated payments, no money moves.
Setup for Claude, ChatGPT, Gemini, Claude Code, Codex, Cursor and VS Code: Connect an AI assistant.
Documentation MCP server
To search the documentation without a MONEI account, use the public documentation MCP server. It needs no sign-in and no key:
https://docs-mcp.monei.com/mcp
claude mcp add --transport http monei-docs https://docs-mcp.monei.com/mcp
Other clients take the usual streamable HTTP entry:
{"mcpServers": {"monei-docs": {"type": "http", "url": "https://docs-mcp.monei.com/mcp"}}}
It is read-only and reaches no account data. Its results are the same records as the search API.
What is available
| Resource | URL |
|---|---|
| MONEI MCP server | https://mcp.monei.com |
| Documentation MCP | https://docs-mcp.monei.com/mcp |
| Page index | https://docs.monei.com/llms.txt |
| Any page as Markdown | https://docs.monei.com/<path>.md |
| Search API | https://docs-search.monei.com |
| Sitemap | https://docs.monei.com/sitemap.xml |
| REST API spec | https://js.monei.com/api/v1/openapi.json |
| GraphQL schema | https://graphql.monei.com/ |
| Postman collection | https://postman.monei.com/ |
Spanish pages are under /es/, with their own /es/llms.txt and /es/sitemap.xml.
robots.txt allows every crawler, and a missing page returns HTTP 404.
Read pages as Markdown
Remove the trailing slash from a page URL and add .md, or send Accept: text/markdown to the
page URL:
curl https://docs.monei.com/payments/refunds.md
curl -H "Accept: text/markdown" https://docs.monei.com/payments/refunds/
llms.txt lists every page with its title, description and Markdown URL.
Search the documentation
The search API needs no key. It is Algolia-compatible, index monei:
curl -G "https://docs-search.monei.com/1/indexes/monei" \
--data-urlencode "query=refund payment" \
--data-urlencode "filters=lang:en AND record_type:docs" \
--data-urlencode "hitsPerPage=5" \
--data-urlencode "attributesToRetrieve=url,hierarchy,content"
Search matches keywords, not meaning: send bizum recurring, not a full question.
Filters
| Attribute | Values |
|---|---|
lang | en, es |
record_type | docs, api, changelog |
type | content, lvl1–lvl6 |
Response shape
Each hit is one section of a page:
url: the page URL with the section anchor. For the Markdown, remove the#anchor, then the trailing slash, then add.md.hierarchy.lvl0–lvl6: the heading path.content: the section text, ornullfor a heading.
Use llms.txt and the .md URLs to read many pages. Do not use search to copy the corpus.
In-browser agents
On every page, a browser agent that supports WebMCP
gets two read-only tools: search_monei_docs and read_monei_docs_page. Nothing to install.
API specifications
The REST API is described by the OpenAPI spec, and the GraphQL API supports introspection. Both need an API key from the dashboard. Use a test key while you build: see Test and go live.
Common questions
Which MCP server should I use?
The MONEI MCP server, when the user has a MONEI account. It also searches the documentation. Use the documentation MCP server only when there is no account.
Which URL should an agent cite?
The page URL with the section anchor, not the .md URL.