Version 1.5 · Release · October 2026

Every target gets what belongs to it.

An additional deployment target now carries its own database — and every view shows the target you selected. Alongside that: the picker carries Sonnet 5.5, Haiku 5.5 and Mistral Large 4, the Quality Gate accepts "this endpoint is open on purpose" as an answer — and the AI notice is back where the law wants it.

Release · October 2026
1.5
5headline features
3new models in the picker
20changes since 1.4
30days of data after going offline

Five highlights

What Version 1.5 adds.

Each card beside the text is a simplified reconstruction of the screen in question.

01Deployment

Every target gets its own database.

An additional target — a staging environment next to production, say — now gets its own database the way the default target always had one, and keeps it for as long as the target exists.

  • Created with the target, kept across every publish, dropped only when it is deleted
  • The database browser, the logs, the metrics and the data export follow along
  • Every view shows the target you selected instead of whichever one came first
  • Additional deployment targets are part of Suite Builder
02Models

Mistral Large 4 takes over from Medium 3.5.

Mistral Large 4 takes the slot Mistral Medium 3.5 held in the model picker — the European house in the list, now in its large size.

  • A project pinned to Medium 3.5 keeps running on it, priced as before
  • The effort control applies here too: Low, Medium, High, per model
  • You choose the model per project, and org rules can narrow the choice
03Quality Gate

An endpoint that is open on purpose.

Before a publish, the gate lists every endpoint reachable without a login. Until now there were two honest answers: secure them, or publish anyway. There is a third now.

  • Tick the ones that belong to the public and the AI marks them in your code
  • The decision lives in your repository, not in a dialog nobody can find again
  • Once the change is committed, the publish restarts by itself
04Models

Sonnet 5.5 and Haiku 5.5 are the new defaults.

Claude Sonnet 5.5 takes over as the default model and Claude Haiku 5.5 as the fast one beside it.

  • Haiku 5.5 costs a tenth of what Haiku 4.5 did per token and reads far more at once
  • A project pinned to one of the previous models keeps running on it
  • Reasoning effort has its own control now, separate from the model
05EU AI Act

The AI notice, where the law wants it.

Article 50(5) asks for the notice to be clear and legible no later than the first interaction. Ours had drifted to the foot of the start screen, below the project tiles and so below the fold on an ordinary display. It travels with the composer again.

  • A first prompt could be sent without the notice ever having been on screen
  • The same paragraph asks for accessibility — the grey reached 1.79:1
  • It is the muted tone the website uses now, at 4.98:1

And much more

Everything else in Version 1.5.

Improvements across the platform — from chat and models to publishing, workspace and account.

Chat & models

  • Reasoning effort has its own control — separate from the model, instead of four letters in every row
  • A start screen that does one thing — the composer first, the rest below it
  • Project rules before the first prompt — the conventions are set before the first line is written
  • A spreadsheet is more than its columns — formulas, sheets and formats come along
  • Two things the chat reported wrong — both corrected

Publishing & hosting

  • Sites now live on airbuild.space — the old xaio.app addresses keep serving, and custom domains are untouched
  • Going offline keeps data for 30 days — database, logs and custom domain stay
  • Two things in the publish dialog — stale findings no longer linger, and the data export is back
  • A downgrade stops extra targets — additional targets are cleaned up properly
  • A notice while an update rolls out — instead of an error that looks like nothing
  • Duplicating a large project works — no more dying at the two-minute mark

Workspace & account

  • Five test runs in one line — one message in the chat instead of five
  • How many tables this project has — now sits where you would look for it
  • Logs, code and API docs follow your theme — no more bright panel in a dark workspace
  • A preview cannot listen in — preview code no longer reaches the workspace
  • A declined top-up no longer blocks — the next one goes through

Ready for Version 1.5?

Active users find everything live in the workspace. New here? Start for free, no credit card.