May 2026: a new engine under the hood

V2's intake parses your prompt into structured intent — project type, vibes, audience, tone, formality, energy, color hints, features. It works in any language. Type in Dutch, Spanish, French, German, Italian, Portuguese, Japanese, English — the engine understands the same way.

The advantage isn't just translation. The intake captures nuance ("sober but warm", "bold dev tool with neon", "calm wellness app") and routes that into the section selection and design tokens, so the output feels matched to what you actually meant — not a generic interpretation.

Typical build now takes 15–30 seconds. The engine routes each prompt through one of three paths: site builds (path 1), foundation-backed app builds (path 2), or full creative builds (path 3) for prompts that don't fit the curated library. Path routing happens automatically; you never pick.

Sections get picked from a curated library, scored against your intent — same prompt twice gives genuinely different (but coherent) results, not the exact same output every time.

For creative sections (path 3), the Deep Engine generates first. If the output fails validation — broken JSX, missing required pieces, length out of range — the engine automatically retries that section through Engine Plus, the strongest lane. You don't ask for it; it just happens when needed.

Premium gets up to 5 rescues per build from the credit pool, Pro gets 10. If you don't need the stronger lane, you don't spend credits there. If you need it, the engine routes there without waiting for you to switch tiers.

Your existing builds, your existing edits, your dashboard, your downloads — all unchanged. V2 is the engine that handles new prompts going forward; everything you've already built keeps working as it was.

Pricing structure is unchanged: Free 1 lifetime build, Starter 4,000 credits, Premium 25,000 credits, Pro 80,000 credits. Cosmetic edits stay free + unlimited on every paid plan. The credit pool is still the only real cap.

The engine itself learns and improves over time — our curated section library grows, foundation templates expand, selection logic gets smarter. For users that means builds become faster, cheaper, and better over time without you doing anything. We aim for the best output at the lowest cost so we can keep delivering more value at fair pricing.

Next milestones we're working on: real interactive preview that works on every device (desktop + mobile), unified chat-style edit/repair surface, and click-to-edit refinement directly from the preview.