Visual editing, a model that reasons, voice as input, private hosting and an engine that runs noticeably faster and more stable. The biggest leap since launch — here are the highlights.
Already on board? Everything is live in your workspace.Every change in the changelog →
Select an element right in the live preview and change it — via settings in the context toolbar or simply by prompt: „Use a typewriter font for the heading.“ Every change is written cleanly into your source code.
XAIO reasons through your request, asks targeted follow-up questions when something is unclear, keeps thinking and then proposes a plan — which you confirm. While building, a live task list shows what's already done.
A new microphone button in the chat and on the start screen records you, transcribes your words and drops the text into the field for review — with a real-time waveform and a blue shimmer.
Pick the right model for the task — from Claude Opus 4.8 through Sonnet 4.6 and DeepSeek V4 Pro to Gemini 3.1 Pro and 3.5 Flash. We optimize model costs in the background and pass the savings on to you — you never have to worry about model prices.
New: dedicated, private hosting per app — noticeably faster than the shared cloud. Choose the right instance from Hobby to Enterprise and change size at any time.
Type @ and the command menu opens: ask a question about the project with @askxaio, get platform answers with @help — or send a teammate a direct message.
A slim bar above the workspace bundles what counts — system notices like new releases and private messages from teammates. Always one at a time, to acknowledge or dismiss.
Over 40 improvements across the platform — from plans and hosting to teamwork and compliance.
Beta 2 also means: less waiting and fewer surprises. We've fundamentally hardened the workspace engine, health checks and publishing.
Active users find everything live in the workspace. New here? Start for free — no credit card, 50 AI Tasks included.
