VeloGen 1.5: Smart Model Routing, Self-Correction & Session Memory

VeloGen 1.5 is live, and it makes every generation smarter and more reliable.

Smart model routing

Not every prompt needs the same horsepower. VeloGen now analyzes each request — length, code context, and complexity signals like auth, payments, or pagination — and automatically routes demanding prompts to a stronger AI model. Simple requests stay fast; complex ones get the depth they need.

Self-correcting code

Every generated answer is now validated server-side: balanced brackets, missing imports, and deprecated API usage (like wix-users when you asked for v2 APIs) are caught automatically. If validation fails, VeloGen regenerates the answer with the problems fed back in — before you ever see it.

Session memory

Follow-ups finally work the way you'd expect. Ask for a repeater, then say "add pagination to that" — VeloGen keeps track of your recent exchanges and resolves references to earlier code.

Also in this release

• Project context: teach VeloGen your real collection names, field keys, and element IDs
• Feedback learning: thumbs-up answers become verified templates for similar future requests
• Live Stripe billing, snippet sharing, and a full visual refresh