Summarize with AI
For a pre-seed AI startup that wants its website built in Next.js, Studio Maydit fits best, because it builds in custom code, works to a fixed scope of three to four weeks, and stays on to design the product once users arrive. Feels Like and Phantom lead the other nine, as the only two that list custom code as their main platform, and both have published AI work. Ranked by fit, the ten studios worth reviewing are Studio Maydit, Feels Like, Phantom, Lazarev, Fantasy, Digidop, Push Refresh, 8020, Edgar Allan, and Refokus. Edgar Allan and Refokus fit least: both are premium Webflow studios serving large brands. To be plain about the stack, none of the nine names Next.js as its framework. Digidop, 8020, Edgar Allan, and Refokus build in Webflow and Push Refresh builds in Framer, so those five would not write your Next.js at all. Lazarev and Fantasy work across mixed tools.
Picture two founders building an AI tool that turns sales call recordings into CRM notes. They have twelve design partners, half an angel round, and a Next.js app started from a template four months ago. There is no website as such. The root page sends logged-out visitors to a sign-in box, and the browser tab still reads Create Next App.
An angel opens the link from their deck on her phone between meetings. She sees a Google sign-in button and nothing else, and she does not sign in. In the founders' analytics, her visit looks like any other.
For this team, Next.js is the right home for the website. The app already runs on it, the engineer founder knows it, and a few public pages beside the product cost less than a second tool with its own login and bill. The job is small: three to six pages that say what the product does, show it working, and collect sign-ups. The hard part is doing it inside a young codebase that changes daily, without slowing the one engineer who also builds the product.
So we put one question to every studio. Could it add those pages to your app this month, cleanly, and leave the founders able to change them the week after?
How we checked each studio for a pre-seed Next.js build
We read each studio's own website and its public directory listings in October 2026. Nothing here was sponsored.
- Who writes the Next.js. We checked the main platform each studio lists. A custom code studio can work in your repository. A Webflow or Framer studio builds somewhere else.
- AI work with young products. We looked for named AI clients and published AI projects. A studio that has explained a young model-based product before learns yours faster.
- A minimum you can see. We noted whether a floor price is public, never the number itself. On angel money, a posted floor lets you skip a call you cannot afford.
- Their team next to yours. We compared headcount with the clients each studio names. A two-person company is a tiny account at a studio that serves Spotify or SAP, and tiny accounts wait their turn.
- The studio's own homepage. We opened each one on a phone with no context. Could we tell in a few seconds what the studio does?
We gave the fifth check the most weight. A studio's homepage is the one project where it chose the words, the stack, and the deadline itself. Whatever it settled for there is what it will settle for on yours.
Every fact in the tables traces to a studio's own site or a public directory profile. Where a studio keeps a number private, its table says Not published rather than a guess.
What goes wrong when a pre-seed AI startup hires for Next.js
Logged-out visitors still hit the sign-in screen. The studio designs a sharp homepage, but the app's middleware guards every route, so anyone without a session gets bounced to the login page. The founders never notice, because they are always signed in. Link previews in chat apps show the login page too. Ask the studio to list which routes are public, and to open each one in a private window on a phone before handoff.
The waitlist form saves nothing. Before seed, design partner sign-ups are the traction slide in the deck. The studio wires the form to a server action, using a key that exists on its own laptop and not in production. The button turns green, a thank-you message appears, and no row is saved. Three weeks of sign-ups vanish before anyone looks. Ask the studio to submit the form on the live site on launch day and show you the new entry.
The build eats your only engineer's week. A studio working inside your app needs access, environment variables, code review, and answers about how sign-in works. At pre-seed all of that comes from one person, who is also shipping the product. If every pull request waits on them, a four-week project takes eight. Ask how many hours a week of your engineer's time the studio expects, and whether a non-technical founder can approve pages from preview links instead.
1. Studio Maydit: A Top-Rated Design Agency for AI Founders
Before seed, a team needs a website that explains the product to an angel in under a minute, and a hard look at the product before early testers stall. Studio Maydit covers both. It is a web and product design studio working with AI founders in the US, UK, and Europe. Sites are built in Framer, Webflow, or custom code, and for an app already on Next.js, custom code is the route that sits beside the product. After launch, the work carries on into product design.
A fixed scope takes three to four weeks, which fits the gap between the first angel check and the wider round. It ends with a diagnosis of what is leaking in the product, so you learn where testers drop off while there are only a few dozen of them. Once the money is in and the team ships every week, a monthly retainer covers new pages, campaigns, and product design, with no long lock-in.
The result to weigh is Dualite, an AI dev tool. Design work supporting a repositioned ICP helped it reach 100,000+ users in seven months. Read how the site and the product moved together in the Dualite case study. The client list also runs to Wave, PixelFlow, Mi-VAD, and 15 other AI and SaaS teams.
| Check | Finding |
|---|---|
| Based in | Remote, serving US / UK / EU |
| Platform depth | Framer, Webflow, and custom code |
| AI-sector proof | Yes. AI-native clients, published outcome on Dualite |
| Pricing | Fixed scope or monthly retainer, quoted per project |
| Team shape | Founder-led, small senior team |
| Best fit | Pre-seed AI founders who want a few clear public pages live in their Next.js app before the raise |
Bring the link you send to angels, and we will open it together, logged out, on a phone. Book a 30-minute call.
2. Feels Like
Feels Like is a Los Angeles studio founded in 2023, nearly as young as some of the companies it would serve. Custom code is its main platform, and it has published AI work, with Suno AI named next to Google, Nike, and LVMH. That pairing is rare in this pool, and the studio sits in its middle tier rather than the premium one.
It does not say which framework it codes in, so confirm Next.js first. Team size and pricing are not published, and a two-person startup may compete with big brands for attention.
| Check | Finding |
|---|---|
| Based in | Los Angeles, USA |
| Founded | 2023 |
| Team size | Not published |
| Primary platform | Custom code |
| AI-sector proof | Yes. Published AI client work |
| Named clients | Google, Nike, LVMH, Suno AI |
| Pricing | Not published |
| Best fit | Pre-seed AI teams with a consumer feel to the product who want a coded site with real character |
3. Phantom
Phantom has studios in London and Auckland, was founded in 2013, and has 51 to 200 people. It builds in custom code and has published AI work for clients such as SAP, Zendesk, Diageo, and the Financial Times. If your pitch only lands when people see the model run, Phantom can build a working demo in the browser.
It is a premium studio at agency scale with pricing not published, so few pre-seed budgets stretch to it. It does not name Next.js, and its work leans toward campaigns.
| Check | Finding |
|---|---|
| Based in | London, UK and Auckland, NZ |
| Founded | 2013 |
| Team size | 51-200 |
| Primary platform | Custom code |
| AI-sector proof | Yes. Published AI client work |
| Named clients | Diageo, SAP, Financial Times, Zendesk |
| Pricing | Not published |
| Best fit | Well-funded pre-seed teams in the UK whose raise depends on a live demo of the model |
4. Lazarev
Lazarev is a San Francisco studio founded in 2015, with 51 to 200 people and published AI work, including Elva. Its process starts with research, which helps when design partners keep asking what a screen means. It posts a minimum, so you can check the fit before a call.
Its platform is mixed, not Next.js, so your engineer may end up coding its designs. It is premium, and research can run longer than a raise allows.
| Check | Finding |
|---|---|
| Based in | San Francisco, USA |
| Founded | 2015 |
| Team size | 51-200 |
| Primary platform | Mixed |
| AI-sector proof | Yes. Published AI client work |
| Named clients | Payoneer, Peel, Elva, Mozayix |
| Pricing | Published minimum |
| Best fit | Pre-seed AI teams whose design partners struggle with the product screens, not just the homepage |
5. Fantasy
Fantasy has worked since 1999 from San Francisco and New York, and it leads with AI strategy next to product and brand work. For founders still arguing about what the product is for, that can settle the story before any page is built.
It names no clients publicly, and its team size and pricing are not published. Its platform is mixed, not Next.js, and at enterprise scale a pre-seed company would likely be its smallest account.
| Check | Finding |
|---|---|
| Based in | San Francisco and New York, USA |
| Founded | 1999 |
| Team size | Not published |
| Primary platform | Mixed |
| AI-sector proof | Yes. Published AI client work |
| Named clients | Not published |
| Pricing | Not published |
| Best fit | Pre-seed teams backed by large funds that need help defining the product story before the build |
6. Digidop
Digidop is a Paris studio founded in 2021, with a team of 1 to 10 and a posted minimum. Its named clients include StreamNative, a data streaming company, so it has explained a technical product before.
It builds in Webflow, not Next.js, so the site would live outside your app on a second tool. Its AI proof is partial, with no AI case study named.
| Check | Finding |
|---|---|
| Based in | Paris, France |
| Founded | 2021 |
| Team size | 1-10 |
| Primary platform | Webflow |
| AI-sector proof | Partial. Software clients, no AI case study named |
| Named clients | TSE Energy, Ramify, StreamNative |
| Pricing | Published minimum |
| Best fit | European pre-seed teams who accept a Webflow site for now and want a small studio with a visible floor |
7. Push Refresh
Push Refresh is a Dallas studio with a team of 1 to 10 and a posted minimum. A small studio with a public floor suits a founder who wants direct contact and a quick, simple build.
It builds in Framer, not Next.js, so none of its work would sit in your repository. Its AI proof is partial, and its founding year is not published.
| Check | Finding |
|---|---|
| Based in | Dallas, USA |
| Founded | Not published |
| Team size | 1-10 |
| Primary platform | Framer |
| AI-sector proof | Partial. Health and services clients, no AI case study named |
| Named clients | SmithRx, Synonym, Northern National |
| Pricing | Published minimum |
| Best fit | US pre-seed founders who decide a Framer site is enough until the round closes |
8. 8020
8020 works from San Francisco and New York and was founded in 2014. Its named clients read like a list of venture-backed software startups, such as Vanta, Pilot.com, and Superlist.
It builds in Webflow, not Next.js, so your app and your site would sit on two stacks. Its AI proof is partial, and its team size and pricing are not published.
| Check | Finding |
|---|---|
| Based in | San Francisco and New York, USA |
| Founded | 2014 |
| Team size | Not published |
| Primary platform | Webflow |
| AI-sector proof | Partial. Startup software clients, no AI case study named |
| Named clients | Wave, Superlist, Pilot.com, Vanta, Circle |
| Pricing | Not published |
| Best fit | Pre-seed teams who want a studio used to venture-backed software and are fine with Webflow |
9. Edgar Allan
Edgar Allan is an Atlanta studio founded in 2014, with 51 to 200 people and clients including Porsche, Duracell, and NCR. It suits large Webflow builds run by a full marketing team.
That is the opposite of a pre-seed company with no marketer. It builds in Webflow, not Next.js, it is a premium studio, pricing is not published, and its AI proof is partial.
| Check | Finding |
|---|---|
| Based in | Atlanta, USA |
| Founded | 2014 |
| Team size | 51-200 |
| Primary platform | Webflow |
| AI-sector proof | Partial. Enterprise clients, no AI case study named |
| Named clients | Porsche, Duracell, NCR |
| Pricing | Not published |
| Best fit | AI companies well past pre-seed that need a large Webflow site run by a marketing team |
10. Refokus
Refokus is a remote studio based in Germany, founded in 2021, with 11 to 50 people. Its named clients include Spotify, BASF, and BCG, brands with strict review processes.
Those are large companies, not founders with a waitlist. It builds in Webflow, not Next.js, it is a premium studio with pricing not published, and its AI proof is partial.
| Check | Finding |
|---|---|
| Based in | Germany, remote |
| Founded | 2021 |
| Team size | 11-50 |
| Primary platform | Webflow |
| AI-sector proof | Partial. Large brand clients, no AI case study named |
| Named clients | Mural, BASF, Spotify, Yahoo, BCG |
| Pricing | Not published |
| Best fit | Later-stage European AI companies planning a big brand site in Webflow |
How to choose between them
Sort by what an angel sees today when she opens your link.
A sign-in box, or a starter page with a default title. You need public pages added inside your app. Studio Maydit, or Feels Like once it confirms it codes in Next.js.
A clear page, but the product only makes sense when it runs. Phantom, if the round can pay for a working demo in the browser.
Pages that exist, yet nobody can say what the product does. That is a story problem first. Lazarev for the product screens, or Fantasy for the strategy.
You have decided the site can live outside the app until seed. Digidop or Push Refresh at a posted minimum, or 8020 for its startup client list. Edgar Allan and Refokus are better kept for a later stage.
One test before you sign. Send the studio your live link and ask what a logged-out visitor sees on a phone in the first five seconds. If it answers without opening the link, keep looking.
