Productivity · SaaS

Magai revenue: a no-code AI platform near $100K MRR

Magai puts many AI models in one subscription. Magai revenue grew to nearly $100K MRR for a founder who built it with no-code tools.

TL;DR: Magai is an all-in-one AI platform that gives users the main text, image and video models in one place for one subscription. Non-technical founder Dustin Stout built it with no-code tools, reached $20K MRR within six months (founder-reported, October 2023) and nearly $100K MRR by February 2026 (founder-reported, Indie Hackers). He says growth came from word of mouth and an email list of nearly 100,000 subscribers, with no paid ads.

Founder: Dustin Stout · By · Updated

Magai at a glance
Revenue Nearly $100K MRRFounder-reported Source, 2026-02
Team size Not disclosed
Pricing One subscription with every model and feature; plans differ only by usage volumeFounder-reported Source, 2026-02
Founded After ChatGPT's launch; exact date not disclosed
Founder Dustin Stout

When ChatGPT came out, Dustin Stout had two failed products behind him, depleted savings and a real chance of losing his home. He could not code. What he did have was more than a decade of blogging and social media, and an email list of nearly 100,000 people. He used no-code tools to build Magai, one place to use the best AI models, and by February 2026 it was close to $100K in monthly recurring revenue (founder-reported, Indie Hackers).

What made a failed actor build an AI platform?

Stout was supposed to be an actor. After auditions dried up in Hollywood he worked as a youth director and then designed t-shirts at a print shop, where he taught himself web design and discovered social media. By 2014 he was a full-time marketing consultant. In 2020 he left the product company he had co-founded, and two failed products followed.

“Rock bottom is a powerful motivator.” — Dustin Stout (founder-reported, Indie Hackers)

Then ChatGPT arrived, and Stout used it obsessively. He kept hitting the same walls: he could not reach the model he wanted, and the interface felt clunky. The wider problem was subscription sprawl, with people paying for several AI tools and juggling API keys. Magai bundles the leading text, image and video models into one interface for one price. The name is “magic” plus “AI”.

How do you build an AI product without writing code?

Stout built Magai before AI coding tools made that easy. The first version took about two months and no code at all (founder-reported, Indie Hackers). He combined Bubble, webhooks, Zapier and other automation tools, working nights and weekends with no funding and no team, and he built only features he personally needed.

Growth broke that foundation. Magai hit Bubble’s ceiling, and his engineers added custom Python and Node.js proxy servers to push it further, until the company had to rebuild the whole application.

StageMagai’s stack
Version 1Bubble, webhooks, Zapier and other no-code automation
Stretching no-codeCustom Python and Node.js proxy servers around Bubble
Version 3 (ready early February 2026)Node.js, Supabase, Vercel, Stripe for payments, ConvertKit for email

Stout says he would launch on no-code again because it let him validate the idea fast with no technical resources. What he would change is timing: plan for the rebuild from the start, because success will require it. In March 2026 he said version 3 was still being rolled out to users slowly.

How did Magai grow with no paid ads?

Stout says Magai grew through word of mouth built on trust, with no paid ads and no growth hacks. That trust took a decade. He learned from the author Michael Hyatt that you can build a platform by consistently giving value, and by launch day he had nearly 100,000 email subscribers and real relationships with experts and influencers who helped spread the word (founder-reported, Indie Hackers).

He also built in public, sharing the messy middle as well as the wins, so early users felt part of the story and became evangelists. The pricing is designed to keep them. Magai sells one subscription with every model, tool and feature included, and plans differ only by usage volume. Magai pays the AI providers based on what users consume and focuses on retention over acquisition.

How did Magai revenue grow to nearly $100K MRR?

Magai went from $0 to $20K MRR in its first six months without marketing and was at $23K MRR in October 2023 (founder-reported, Indie Hackers). In February 2026 Stout said it was right on the cusp of $100K MRR (founder-reported, Indie Hackers). Total revenue and team size are not disclosed.

WhenMilestoneSource
After ChatGPT’s releaseFirst version built with no code in about two monthsIndie Hackers
First six months$0 to $20K MRR without paid marketingIndie Hackers
October 2023$23K MRRIndie Hackers
February 2026Nearly $100K MRR; fully custom version 3 readyIndie Hackers

The pressure behind those numbers was personal. Every decision carried weight because failure meant losing his home, and Stout says that forced a discipline the company still runs on: no vanity metrics, and every feature has to matter.

How is Magai different from TypingMind and Merlin?

All three sell a better way to use the same frontier models. TypingMind gives users a polished chat interface but asks them to bring their own API keys, and it launched with one-time licences. Magai takes the opposite approach: no API keys at all, one monthly subscription, and Magai handles every provider relationship and bill. Merlin puts AI inside the browser, where people already work.

The difference is who carries the complexity. TypingMind’s buyers accept setup in exchange for control over model costs. Magai’s buyers pay for convenience, which suits Stout’s target of everyday workers rather than developers, but it also means Magai’s margin depends on what its users consume. Its distribution is different too: TypingMind and Magai both launched to audiences their founders already had, while Magai’s came from a decade of blogging and email rather than X.

For a standalone AI answer engine that raised venture money, see Perplexity.

Motion bundles AI agents for small businesses rather than giving individuals access to many models.

Newer non-coder founders build with AI app builders instead of no-code tools, such as Plinq on Lovable and Payout, a vibe-coded class action app.

What can founders learn from Magai’s story?

  • Build the audience before the product. An email list of nearly 100,000 people is why “no marketing” worked.
  • Make pricing a non-decision. One subscription with every feature removes the hardest choice for a new user.
  • Validate on no-code, but plan the rebuild. Bubble got Magai to market in two months; outgrowing it was slow and expensive.
  • Build in public. Sharing the honest progress turned early users into the product’s marketing.

Frequently asked questions

How much revenue does Magai make?

Nearly $100K MRR as of February 2026 (founder-reported, Indie Hackers). In October 2023 it was at $23K MRR. Total revenue is not disclosed.

Who founded Magai?

Dustin Stout, a former actor and marketing consultant who taught himself web design. He built Magai alone, with no funding, after two failed products (founder-reported, Indie Hackers).

How does Magai pricing work?

One subscription includes all AI models, tools and features. Plans differ only by usage volume, and Magai pays the AI providers based on what users consume (founder-reported, Indie Hackers).

Was Magai built with no-code?

Yes. The first version was built in about two months with Bubble, webhooks and Zapier. The current version 3 is a full rebuild on Node.js, Supabase and Vercel (founder-reported, Indie Hackers).

Sources

  1. From broke to $100k MRR in four years – Indie Hackers, 2026-02
  2. Hitting $20k MRR in 6 months without marketing – Indie Hackers, 2023-10

Labels: Founder-reported = stated by the founder or company; Verified = backed by a payment-processor screenshot, TrustMRR or a public platform count; Media-reported = published by a news outlet or third party. "Implied" monthly figures are annual figures divided by 12. We never estimate numbers ourselves.

Last updated: