How to pick the prototyping tool

Pick on a real task, and pick by what the prototype is for.

First decide what job the prototype does. There are three, and they want different tools:

Validation — put something in front of a user and learn. Needs realism and speed, not code quality.
Alignment — show the team what you mean instead of writing it. Needs to be shareable and editable fast.
Handoff — become the starting point for the real build. Needs exportable, readable code.

Most PMs are doing the first two. If you’re choosing a tool for handoff, you’re probably solving an engineering problem you weren’t asked to solve.

The two families:

Design-first (Figma Make and its peers): starts from your design system, produces something that looks like your product. Best for alignment, weaker when the prototype needs real logic or data.
App-first (Lovable, Bolt, v0, Replit): produces a working app with state, forms, and a database. Best for validation, because users can do real things in it. Looks generic unless you push it.

Pick the family by which failure hurts more: a prototype that looks off-brand, or one that can’t do anything.

Then the test. Take one real feature idea. Build it in two tools in one afternoon. Score on:

Time to first clickable version. Under an hour, or it’s not a prototyping tool.
Iteration. Ask for three changes a user might request. Does it make them, or break what worked?
Deploy. Is there a URL you can send to someone by the end? If not, it’s a demo tool, not a prototype tool.
Data. Can it hold real-ish data — a list, a login, a saved state? Validation prototypes need this; alignment ones don’t.
Escape hatch. Can you get the code out, or edit it directly when the tool gets stuck? It will get stuck.
Cost per experiment. Credit-based pricing punishes iteration, which is the whole point.

Two things nobody tells you:

The tools are converging fast. Six months from now the differences will be smaller. Don’t over-optimize; pick one and get fluent.
Once you’re comfortable with an agentic coding tool, it quietly becomes your prototyping tool too — it builds the same thing, with more control and less hand-holding. Many people end up starting on Lovable and graduating to Claude Code or Cursor within a few months. That’s a normal path, not a failure of the first tool.

For you specifically: you have a design background, so the app-first family’s generic look will bother you — but you’re validating ideas, so you need the logic. Start app-first, fix the look yourself. That combination is rarer than either alone.