Quickstart

Updated 2026-10-18

Connect an AI assistant to the statistics, check it works, then use the data in code and put a game on a page. No account and no API key are needed for any step.

1. Connect your assistant

The hosted server works in every client that supports remote MCP servers. Claude Code is one command:

claude mcp add --transport http --scope user dyscalculia https://mcp.dyscalculia.io/mcp

For other clients use the same URL: https://mcp.dyscalculia.io/mcp. Step-by-step pages: Claude Code, Claude (claude.ai and Desktop), Claude Desktop (local), ChatGPT, Codex, Cursor, VS Code, Gemini CLI, Any other client.

2. Ask a question

Using Dyscalculia.io, what share of school-age children have dyscalculia, and what is the source?

The answer should be about 3 to 6 percent, with the study it comes from and its year, read by the get_prevalence tool.

3. Check without a client

curl -s https://mcp.dyscalculia.io/mcp \
  -H 'content-type: application/json' \
  -H 'accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

The address https://mcp.dyscalculia.io/health answers ok. The server is stateless and accepts only POST /mcp, so a browser visit to /mcp shows 405.

4. Use the data directly

npm install dyscalculia-data

import { prevalenceFor, getReference } from "dyscalculia-data";
prevalenceFor({ study: "israel-1996" });  // 6.5 percent, 3,029 children, 1996
getReference("grosstsur1996");

Files, columns and examples in curl, Python and R are on the dataset page.

5. Put a game on a page

<div data-dysc-game="dots"></div>
<script async src="https://dyscalculia.io/embed.js"></script>

Options, languages, chart embeds, platform steps and the credit-line terms are on the embed page.