FinBaze for Platforms

SDKs

Everything on this site is also available to AI tools, in five ways. All of them are public and read-only.

WhatWhereUse it for
Docs MCP serverhttps://mcp.finbaze.com/docs/mcpAn agent that looks things up while it writes code
llms.txt/llms.txtA short map of the site with the hostnames and auth in one screen
llms-full.txt/llms-full.txtEvery page as one Markdown file, for tools that load the whole thing
.md per page/quickstart.md, /guides/banks.md, …One page as Markdown
Copy for AIThe button on every pageThe same Markdown, on your clipboard
Agent skillAgent skill"Integrate FinBaze embedded components into a Next.js app"

Docs MCP#

Add the server to your agent. It needs no sign-in and holds no session: each request stands alone.

{ "mcpServers": { "finbaze-docs": { "url": "https://mcp.finbaze.com/docs/mcp" } } }

Claude Code: claude mcp add --transport http finbaze-docs https://mcp.finbaze.com/docs/mcp.

ToolInputReturns
search_docsquery, section?The top 8 pages with a snippet and the URL
get_pagepathThe page as Markdown
get_graphql_typenameThe SDL of a type, its description and the operations that use it
get_operationnameThe operation: signature, scope and feature it needs, a document with example variables, and a request snippet in Node, Kotlin, PHP and Python
list_components, get_componentkey?The component catalogue, or one component with its options, features, events and requirements
get_webhook_eventtypeThe event's description and a sample delivery
get_error_codecodeWhat an error code means and what to do

The index is built with the site, from the same Markdown and the same generated references, so the server never answers from an older version than the page you are reading. It lists the operations a platform key or token can reach, not the whole GraphQL schema of FinBaze.

The request snippets call the HTTP API directly (/v2/oauth2/token, then /graphql). Typed SDK methods for Kotlin, PHP and Python are generated from the same documents and arrive with those SDKs.

The server is rate limited per client. Do not point a loop at it; fetch llms-full.txt instead.

llms.txt and Markdown#

llms.txt follows the llms.txt convention: a title, a summary, then links. llms-full.txt is every page in order, each preceded by an HTML comment with its path. Append .md to any page address for its Markdown (/guides/pos.md), or press Copy for AI on the page.