Changelog

What changed in AiBrush Studio, newest first. Every entry below describes something that is live in the product today; work in progress is not listed until it ships.

  1. · v3.5.4

    Review timeline, AI revisions, and credit holds

    Ad Studio gains a review timeline with timestamped notes and version history, and credits are now held while a generation runs and settled when it finishes.

    In Ad Studio you can scrub an ad, select ranges, and leave timestamped notes with attachments, then submit them as one feedback batch that can drive an AI revision. Every compose, revision, and restore is kept as a version you can switch back to instantly. Across the studio, your balance is reserved when a generation is queued, charged only when it succeeds, and returned when it fails, and account emails now link to real pages.

  2. · v3.5.3

    AiBrush Video Model and faster video

    A first-party video model with synchronized audio, text- and image-to-video with keyframe control, and much faster clip delivery.

    The new AiBrush Video Model produces clips of one to ten seconds at around 1080p with a matching audio track and sits alongside the existing video models in the Video Studio. Generate from a prompt, pin the first frame from an image, or condition on first and last keyframes across eight aspect ratios. Clip delivery is several times faster than before, every video model gets the same audio on/off switch, and renewals, upgrades, and cancellations are reflected in your plan immediately.

  4. · v3.5.1

    The new AiBrush Studio

    A rebuilt web app with dedicated creation tools, a unified asset library, per-model credit costs shown before you generate, and full English, Arabic, and Japanese localization.

    AiBrush Studio was rebuilt from the ground up for a faster, cleaner experience. Every creation workflow has its own page: image generation, product and model photography, magic edit, multi-angle shots, upscaling, portrait retouch, outfit, face, and character swap, the Video Studio, voiceover, voice cloning, voice changing, music, and sound effects. A single Assets library gathers everything you make, with drag-and-drop uploads and one-click retry on failed generations. Credit cost is computed per model and shown before you submit, and the whole studio is available in English, Arabic, and Japanese with locale-aware links.

How to read this page

Each release has a stable link you can share, a machine-readable date, and a short summary of what changed and why it matters. The dates are the days those releases went live.

This page lists shipped changes only. Roadmap items and experiments stay off the list until they are available to everyone, so anything you read here you can use today.