EN
Back to the archive

The archive · AI & Models · Product decision · 2023–2025

TypingMind: a solo dev's ChatGPT UI hits $145K MRR with no funding

A weekend ChatGPT-UI wrapper becomes a $1M+ ARR solo business, bootstrapped from day one.

TypingMind

The betThat users would pay for a better chat interface on top of ChatGPT — search, plugins, multi-model freedom — instead of using OpenAI's default UI.Scaling

What the business is

A paid chat UI for LLMs that wraps ChatGPT and other models with conversation search, plugins, prompts, and team deployment options.

Starting capitalNone; solo bootstrapped from prior indie product income

How it started

On March 1, 2023, OpenAI announced the ChatGPT API. Vietnamese indie hacker Tony Dinh — two years into solo building after quitting his job — decided the default chat UI was too limited and registered typingmind.com the next day. He released the first version on March 6 and announced it on Twitter.

What happened

Sales took off immediately: $1K on day one, $22K in the first week, and #1 Product of the Day on Product Hunt with 4,000+ paying users within three months. He kept adding models and features, introduced a Team version for companies building internal AI platforms, and closed a 3,000-seat enterprise deal. By November 2024 the product had passed $1M in trailing revenue.

How it ended up

Still running and growing solo-led: about $130–160K a month as of October 2025, with recurring B2B revenue above 50% and the founder financially independent.

Background

On March 1, 2023, OpenAI announced the ChatGPT API. Tony Dinh, a solo indie hacker in Vietnam who had quit his job two years earlier, saw a gap: the official ChatGPT interface had no conversation search, no plugin ecosystem, and no way to use multiple models. He built a better UI for himself over a weekend, registered typingmind.com the next day, and released it publicly on March 6.

The product sold almost immediately — $1K on day one, $22K in the first seven days — and was #1 Product of the Day on Product Hunt within months, with 4,000+ paying users. Dinh priced it as a one-time license and kept raising the price as he added features, then added a subscription Team version for companies building internal AI platforms.

By November 2024, TypingMind had passed $1M in trailing revenue, including a 3,000-seat enterprise contract. As of October 2025 it was making roughly $130–160K a month, with recurring B2B revenue above 50%, still bootstrapped and mostly run by its solo founder.

What has to be true

  • The bet was that the interface layer around a fast-moving model API was itself a product worth paying for, not just an afterthought.
  • Shipped in five days, before clones and official improvements arrived, converting a temporary window into a durable audience.
  • A pricing ladder (one-time license to team subscription to custom deployment) let one product serve consumers and enterprises without VC.
  • Organic word of mouth — positive Reddit mentions and a large Twitter following — replaced paid marketing, which he found uneconomical.

What can be applied

When a big platform opens an API, speed to market matters more than polish: a solo founder who shipped in five days captured demand that a larger team might have over-engineered away.

Aftermath

As of October 2025, TypingMind continues to ship under Dinh's solo leadership, with a small team supporting sales and support. Monthly revenue is roughly $145K with the B2B Team version exceeding half of it, and the founder reports financial independence while still working about four hours a day on the product.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases