Live Types

How it works

1. Paste one or more real API response samples (not a hand-written schema). Live Types diffs the samples against each other, so a field that's a string in one sample and null in another becomes string | null instead of silently picking one.
2. It infers a TypeScript type and a matching Zod schema from the merged shape — runtime validation and compile-time types from the same source, so they can't drift apart from each other.
3. Save the inferred shape as a snapshot. Next time you paste a fresh response from the same endpoint, Live Types diffs it against the snapshot and tells you exactly which fields were added, removed, or changed type — the contract-drift check most teams only discover in production.

Why not just use quicktype

quicktype is the better choice if you need many output languages from a single schema. Live Types is narrower on purpose: multi-sample inference (so union types come from evidence, not a guess), a generated Zod validator alongside the type, and snapshot diffing built in — for the specific job of tracking whether a third-party or internal API is still shaped the way your code assumes it is. Full quicktype vs Live Types comparison.

Frequently asked questions

Does anything I paste leave my browser?

No. Inference, validation, and diffing all run client-side. There is no backend and no account — see the privacy policy for what the ad and analytics scripts do collect.

Can I check for drift automatically, not just when I paste a new sample?

Not yet inside the browser tool. It has no server, so it can't poll an API in the background. The roadmap includes an exportable snapshot format a CI job can diff against on every run, so drift detection becomes part of your pipeline instead of something you remember to re-check manually.

Who maintains this?

One developer, not a company — see About & Contact.