Ultimate Guide to AI Prototyping Tools (Lovable, Bolt, Replit, v0)

Here's what you need to know to use these tools in a forward-thinking product development process

The most significant new AI tool to release for PMs in the last 12 months has been the rise of AI prototyping tools:

• Bolt has grown from 0 to $40M ARR in 6 months (CEO on the podcast this week)

• Lovable has surged from 0 to $17M ARR in just 3 months

• Vercel (maker of v0) is valued at $3.25B

• Replit is valued at $1.16B

It's quietly redefining what it means to be a PM in forward-thinking orgs.

---

Introducing Colin Matthews

My most popular podcast to date was with Colin Matthews. So we knew we had to team up again—this time, to write the definitive newsletter on the topic.

Colin teaches a popular four-week course on AI Prototyping for PMs, where you'll build your first AI prototype even with zero coding skills. You’ll build 5+ prototypes and master debugging strategies to ship faster.

Check out Colin's course on Maven and use code AAKASHxMAVEN to get $100 off.

---

Today’s Post

We have endeavored to put together the web’s most tactical guide on AI prototyping for PMs, designers, and engineers yet:

1. How to Understand the Tools Out There

2. How to Use these in Your Product Development Process

3. 🔒 Tutorials of The Top 4 Tools

   • v0

   • Bolt

   • Replit

   • Lovable

4. 🔒 How PMs Should Be Using AI Prototyping

5. 🔒 A Guide to Debugging

6. 🔒 Prototype to Scale

---

1. How to Understand the Tools Out There

We are talking about cloud development environments (CDEs) when we say AI prototyping tools. CDEs have 4 key players right now:

All four tools are great - and they will likely build to parity over time. But to describe when to use each of the 4 major players in a sentence, we would say:

• v0 - beautiful front-end design by default

• Bolt - quick prototypes with flexible iteration power to expand

• Lovable - easy to use for non-technical users

• Replit - products that require working with persistent data or internal tools

Here’s the layer deeper on each.

Overview - v0

Vercel’s v0 shines at crafting polished front-end prototypes fast, leveraging its $3.25B valuation to prioritize UI finesse over Bolt’s raw speed.

PMs love it for stakeholder demos or cobbling together pre-built components from its community, though it leans on third-party integrations like Supabase for databases, unlike Replit’s built-in solution.

It can run server-side functions but requires hosting with Vercel for deployment.
Most people don’t use v0 for servers/dbs, but it is now capable of handling them.

Overview - Bolt

Bolt’s strength lies in blazing-fast prototyping, hitting $40M ARR in 4.5 months by turning screenshots into full UIs instantly, outrunning v0’s design polish.

As the CEO told us, Bolt differentiates from the other 3 tools by having the entire environment run in the browser, instead of on VMs (virtual machines). It’s perfect for PMs iterating in real-time—like Colin’s Apollo scheduler demo in 10 minutes—though it needs Supabase for persistence, unlike Replit.

It has more flexibility but fewer guardrails, leaving the user to pick the best tools and libraries for their needs.

Overview - Replit

Replit, valued at $1.16B, dominates with full-stack power, offering built-in databases and auth that Bolt and v0 can’t match without external tools.

It’s the pick for PMs building lasting tools—like Colin’s referral app with Stripe integration—deployable in minutes.

It sacrifices Bolt’s speed and Lovable’s design fidelity for depth, ideal when prototypes need to scale into real products.

Overview - Lovable

Lovable’s $17M ARR in 3 months reflects its easy of use and low technical barriers.

Its selector tool and Supabase sync make it friendlier for non-technical PMs than Replit’s code-heavy approach.

It’s less nimble than v0 for quick UIs and lacks Replit’s native database, prioritizing integration over raw prototyping pace.

Adjacent Tools

There are 2 adjacent sets of tools out there:

• Chatbots: Claude, ChatGPT

• Coding AI Tools: Cursor, Windsurf, GitHub Copilot, Zed

They are not exactly AI prototyping tools but they are very powerful. So let’s talk about them briefly.

Chatbots

Think of chatbots overall as good for one time prototypes, but if you want to iterate and improve, use a cloud development environment.

There are two top tools:

1. ChatGPT will write code for you pretty well. And also just released the functionality to run basic web apps on its own.

2. Claude’s Artifacts system allows you to run the code in Claude itself. But, right now, you can’t make changes to the code directly in Claude. And you tend to hit usage limits pretty fast, without a way to pay more to use more like the cloud development environments. It’s best for single file prototypes.

Coding AI Tools

There are a bunch of really powerful coding AI tools out there (Cursor, Windsurf, Github Copilot, and Zed are leaders of the pack). These are best for people who know how to code, and want to ship the app to production.

These tools can do a lot more than cloud development environments, but they require technical literacy and aren’t as fast for simple prototyping tasks.

Many times, a strong workflow these days is to start in a cloud development environment, and then migrate into a coding AI tool.

For example, Colin started a presentation app that he built using Lovable. He synced that code to Github. Then he used Cursor, the coding AI tool, to build on top of the app.

Cursor, in particular, is growing really fast and excels at resolving bugs and troubleshooting issues. So once you get over the technical interface, even non-technical people teach themselves to use it.

If you’re a startup PM, you may want to go that route. But today’s post will focus on cloud development environments.

---

2. How to Use These in Your Product Development Process

So with these new, powerful tools, how should you be using them?

Here’s how this plays out in product development:

The Old Way

Here’s what most product development lifecycles look like:

1. Ideation: Most teams barely prototype at the idea stage. A rare few exceptional designers and PMs do (~5%)

2. Planning: Here, more teams use prototypes, but it still is an exception few (~15%), while sketches and mockups are much more common (>75%)

3. Discovery: For many in feature factories, this step is skipped completely. But in more empowered companies, many teams would test prototypes in the discovery phase (~50%)

4. PM Handoff: It's rare for PMs to include prototypes in their PRDs—often met with side-eyes from designers who prefer sketches or plain descriptions (5%).

5. Design Exploration: This is the most common spot where prototypes come into the picture, and most product teams looking to reduce risk of big features prototype (~75%)

6. Engineering Start: Because many prototypes are not engineered, engineers would start from scratch with a final design.

This has been steady for the past several years.

The New Way

But something is changing amongst forward-thinking teams. PMs are moving beyond documents and closer to the “bare metal” of the pixels that actually define a product.

Prototypes have become the new way to communicate your ideas - at all stages:

1. Ideation: They’re using prototypes to work out product problems.

2. Planning: They’re pairing roadmap discussions, pre-PRD, with prototypes.

3. Discovery: They’re putting these prototypes in front of real potential customers to explore solution spaces.

4. PM Handoff: Then they’re attaching working prototypes to their PRDs—turning ideas into clickable clarity.

5. Design Exploration: Designers are going from a PM-fidelity prototype to a design-level prototype (often in Figma) that fits into the design systems and goals of the team.

6. Engineering Start: Engineers may even start with the latest prototype from a tool to get a headstart on their code. Then they’ll use AI coding tools like Cursor or Windsurf to get it ready for production.

In other words: All 3 of PM, design, and engineering are now using these prototyping tools.

(It does bring up an interesting question: where does the role of the PM end and where does the role of the designer begin? We’ll talk about that more in section 4.)

---

3. Tutorials of Our Favorite Tools

Now, let’s get into how to very tactically use each one of these tools: