Public beta · API platform
Weft reads your OpenAPI file or your TypeScript types, then serves typed endpoints, hosted docs and a client SDK from the same source. Every push deploys. Every deploy rolls back in one click.
Median response time over the last 30 days: 38 ms across 4 regions.
$ weft init --from openapi.yaml✓ 14 operations read✓ Typed SDK written to ./sdk (3 files)$ weft deploy --env production→ Building for 4 regions✓ Live at https://api.example.com/v1✓ p50 38 ms · p99 121 ms$ Weft works from the spec you already keep. You do not write routes, hand-roll clients or wire up docs by hand.
The SDK comes from the same schema that runs your endpoints. Rename a field and the build fails in your editor, not in a customer's checkout.
const order = await weft.orders.get("ord_72hK")
order.total // number, checked against the live schemaEach deploy keeps its previous build warm for 14 days. Roll back one region or the whole fleet from the dashboard, or run weft rollback.
Limits sit next to each route. Responses carry the remaining quota in headers, so clients slow down before they meet a 429.
How a request moves
Short pages with working code. Every example runs against the public beta.
Price by seats and regions, not by every request. The full comparison is on the pricing page.
£0
One seat, one region, for side projects.
See HobbyMost teams start here
£29 per seat a month
Four regions, 10 million requests included, 30 days of logs.
Request beta access£249 a month
Every region, 50 million requests, a 99.95% uptime target.
See Scale“We went from a sprint of SDK upkeep each quarter to an afternoon, and the docs stopped drifting.”
Or write to hello@example.com. We reply within one working day.