Get It Live · trust filter

Know the problem. Define success. Pick the AI. Ship.

AI did not invent product thinking. It removed the need for a team between an insight and a live product. Use this site to judge which AI tools deserve time, effort, and money.

The path we teach

How we work

01

Know and love the problem

Start with a problem you understand. Taste and judgement stay with you.

02

Define success criteria

If you can say what good looks like, you are already more than halfway.

03

Choose AI tools that match

This directory. Pick tools that fit the problem.

04

Get it live

From a working build to people using it. Shipping is the point.

The directory covers step three.

How we score

Four sections. Not a catalogue of every model. Tools that can help one person who knows a problem get something live.

Listings

All listings

Each listing is a judgement against time to live, money to useful, effort to own, problem-fit, and lock-in. Example cards are marked. Do not cite examples as product advice.

Demos Yes, with caveats

Bolt (StackBlitz)

StackBlitz's AI full-stack builder at bolt.new that turns prompts into runnable websites, web apps, and Expo mobile projects in the browser, with Bolt Cloud hosting, GitHub sync, and ZIP export.

Build with agents Yes, with caveats

Claude Code

Anthropic's coding agent (terminal, editor extensions, Desktop Code tab, and cloud sessions) for reading a codebase, editing files, running commands, opening pull requests, and carrying development work through to release.

Build with agents Yes, with caveats

Cursor

Coding platform (desktop editor, command line, web, and Cloud Agents) for product and engineering teams to explore a codebase, plan and implement changes, open pull requests you can review, and review those requests with Bugbot.

Build with agents Yes, with caveats

Devin

Cognition’s AI software engineering platform (Devin Cloud web agent, Devin Desktop, CLI, Slack/Teams, API, Review, and Automations) for planning, implementing, testing, and shipping production changes across real repositories in the product development lifecycle.

UX research Yes, with caveats

Dovetail

Customer intelligence platform that transcribes, tags and summarises interviews, usability sessions, surveys and feedback, with AI Chat on Free and Agents and Digital Twins on Enterprise. It analyses evidence you bring; it does not recruit participants or run tests.

Build with agents Yes Example only

Example Agent Loop

Example listing for Build with agents. Not a review of a real product.

Demos Yes, with caveats Example only

Example Demo Studio

Example of a prompt-to-demo listing. Not a review of a real product.

Design systems No Example only

Example Token Bridge

Example of a design-system listing. Not a review of a real product.

Demos Yes, with caveats

Figma Make

Figma's AI prompt-to-app tool that turns natural language and existing Figma designs into interactive prototypes and web apps, with Make kits for design-system context, share/publish URLs, ZIP and GitHub export, and (closed beta) local-codebase PR handoff.

Demos Yes, with caveats

Google AI Studio

Google's browser tool for trying Gemini, and for building web or Android apps from plain-language prompts, with Firebase, GitHub sync, Cloud Run publish, and a ZIP download.

UX research Yes, with caveats

Great Question

Research platform for interviews, surveys, and unmoderated tests on Figma prototypes or live apps, with a list of your participants, a pay-as-you-go panel, and a connection for coding assistants (MCP). You can buy a seat without a sales call, at $129 a month, and there is no free plan.

Demos Yes, with caveats

Lovable

AI full-stack app builder that turns natural-language prompts into editable React web apps (UI, backend, auth, publish), with GitHub/GitLab sync for engineering handoff.

Demos Yes, with caveats

Lovable vs v0 vs Bolt.new: which to pick for a demo by Friday

Three chat-to-app builders that all get a clickable, hosted demo live in days. Pick Lovable for a full-stack demo a non-engineer can drive, v0 if the demo must look like your product or land in your Next.js repo, and Bolt if you want the largest free allowance and a Figma frame as the starting point.

UX research Yes

Lyssna

Self-serve UX research platform (formerly UsabilityHub) for unmoderated tests, surveys and interviews, with a pay-per-minute participant panel open to every plan, including Free.

Demos Yes, with caveats

Magic Patterns

AI design tool for product teams that turns prompts, product briefs, screenshots, and imported design systems into interactive React prototypes, with a shared canvas, publish URLs, Figma export, GitHub sync, and a handoff into Cursor or Claude Code.

UX research Yes, with caveats

Maze

Self-serve user research platform for unmoderated prototype and usability tests, surveys, card sorts and interviews, with a pay-per-response participant panel and an Enterprise-only AI moderator.

Design systems Yes, with caveats

Relume

Marketing-site component system for Figma, Webflow, and React, maintained by people. AI Site Builder assembles pages, Relume Publish hosts them (beta), and a library connection (MCP) lets coding agents use the real components instead of a made-up interface.

Demos Yes, with caveats

Replit Agent

Replit's autonomous AI builder that turns natural-language prompts into full-stack apps (and designs, slides, and more) inside the Replit cloud IDE, with Plan/Build modes, Design Canvas, built-in auth/database/hosting, checkpoints, and GitHub handoff.

Design systems Yes, with caveats

Specify

Design token engine that gathers tokens and assets from Figma and Tokens Studio, stores them in Specify's token format (SDTF), and sends production-ready output through parsers, GitHub pull requests, a command-line tool, a developer kit (SDK), and an HTTP API.

UX research No

Sprig

Enterprise survey platform with AI agents that design, field and synthesise studies, plus in-product surveys you can drop into AI-built prototypes. Sales-led pricing and enterprise-only panels make it a poor fit for a solo builder.

Design systems Yes, with caveats

Supernova

Design system platform that manages your tokens, components, and docs, automates design-to-code pipelines, and gives coding agents the same source through MCP, a standard connection for those agents.

Demos Yes, with caveats

v0 by Vercel

Vercel's AI builder that turns prompts (and mockups) into React/Next.js UIs on shadcn/ui, with live preview, GitHub PR handoff, and one-click Vercel deploy.

Who this is for

  • Product managers, product designers, marketers, and founders.
  • Anyone with an idea who can define success and will ship.
  • People choosing a few tools for a product they will ship.

Who this is not for

  • Anyone shopping a long catalogue of AI tools.
  • Teams looking for a playbook that hands the work to engineering.
  • Tools that only make ads or short videos.