Dyscalculia.io developer docs
Three free, open-source building blocks from Dyscalculia.io: the dyscalculia statistics as data (CSV, JSON and an npm package), an MCP server that lets AI assistants such as Claude, ChatGPT, Codex and Cursor look the figures up with their sources, and 5 number practice games you can embed in 6 languages. 7 population studies and 93 references, each with its year and source.
Start here
| I want to… | Go to |
|---|---|
| Let an AI assistant look up dyscalculia figures with sources | MCP server |
| Download the statistics as CSV or JSON, or use them in code | Statistics dataset |
| Put a free number game on my website | Embed the games |
| Try it in two minutes | Quickstart |
| Read the statistics as a page | Dyscalculia statistics |
What is published
| Package | Version | What it is | Source |
|---|---|---|---|
dyscalculia-mcp | 0.1.0 | Read-only MCP server: 6 tools, stdio and hosted HTTP | GitHub |
dyscalculia-data | 0.1.0 | The statistics, references and screening-game item sets as JSON and CSV with small helpers | GitHub |
dyscalculia-embed | 0.1.0 | The free number practice games as a web component and React wrapper | GitHub |
What the figures are, and are not
The statistics are prevalence estimates from published studies and official counts, each with the year of the study and a reference you can check. There is no worldwide pooled meta-analysis of dyscalculia prevalence, so the overall figure is a range across population studies (3 to 6 percent; the German clinical guideline gives 3 to 7), not one number. Studies use different criteria and ages, so they are not directly comparable.
Licenses
- Code (MCP server, data package, embed package, these docs): MIT.
- Statistics compilation (selection, tables, charts, descriptions): CC BY 4.0. Credit Dyscalculia.io and link the statistics page. Each figure belongs to the source it cites.