Publishing an open dataset on GitHub is easy. Getting developers to use it is a different problem. Here is what I learned shipping ReceiptEdit's 2026 US sales tax dataset as a free REST API.
1. No auth. No sign-up. Ever.
Every API-key ceremony is a step where trial users bounce. The dataset is public. https://receiptedit.com/api/sales-tax/california — that is the entire contract.
2. Match the data shape to how it will actually be consumed
90% of consumers ask "What's the sales tax for X state?" So the top-level endpoint is /api/sales-tax/:state — one state per response. When someone needs all 50 for analysis, the CSV/JSON files sit next to the API at github.com/receiptedit/us-sales-tax-2026.
3. CORS on. Cache long.
Access-Control-Allow-Origin: *
Cache-Control: public, max-age=86400, s-maxage=604800
4. Ship an embed alongside the API
<iframe src="https://receiptedit.com/api/embed/sales-tax/california" width="380" height="540" style="border:0" loading="lazy"></iframe>
Bloggers on Medium/WordPress cannot install an npm package. They can paste an iframe.
5. License permissively
Real MIT. Attribution appreciated, never required.
Try it
- API: https://receiptedit.com/api/sales-tax
- Source: https://github.com/receiptedit/us-sales-tax-2026
- Embed: https://receiptedit.com/api/embed/sales-tax/california
- Docs: https://receiptedit.com/developers
- Product: https://receiptedit.com
Happy to talk API-design tradeoffs in comments.
Top comments (0)