Buildpad: Revenue, team and how it was built
Buildpad (AI co-founder app that researches demand and plans a product, Romania): US$6K MRR, July 2025. Evidence grade C: the founder's own figures, not verified by anyone else. Team: 2 co-founders, the brothers Felix and David Heikka; a team planned for 2026. Launched 19 August 2024; still running, revenue growing.Added 26 October 2026
- Outcome
- US$6K MRR, July 2025Revenue growing
- Evidence
- C the founder's own figures, not verified by anyone else
- Launch
- 19 August 2024
- Status
- Still running
- Team
- 2 co-founders, the brothers Felix and David Heikka; a team planned for 2026
- Outside funding
- None found in the sources
- Main channel
- Reddit and communities
Buildpad is a web app that calls itself an AI co-founder: an AI assistant that leads a person through the steps of starting a business, asks questions, and pushes back on weak ideas, instead of only answering what it is asked. Its main feature uses AI agents, programs that run a large language model in a loop with tools such as web search, to read Reddit and X discussions, app-store reviews and market reports and to report whether people really have the problem the founder wants to solve. Users pay with credits, units that the app subtracts for each AI action; research uses more of them than a short chat message.
Two Swedish brothers built it. David Heikka had co-founded an earlier software business and learned to code in 2023; Felix Heikka left the Swedish army in March 2024 and joined him in Budapest. After an AI lead-form product and a dating-profile joke app failed, they posted a survey in founder subreddits in July 2024, released a first version on 19 August 2024 and launched on Product Hunt in October 2024, which brought the first paying customers. Felix reported the MRR, monthly recurring revenue, the subscription income that repeats each month, in Hacker News posts: US$2,300 in January 2025, US$2,700 in February, US$3,600 in April and US$6,000 on 1 July 2025, with US$33,000 of total revenue by June 2025. The company behind the app is BUILDPAD S.R.L., registered in Romania.
Every revenue number on this page is the founder's own statement on Hacker News, grade C: the figures rise steadily across seven dates, but no payment processor or journalist has confirmed them, and no revenue figure after July 2025 was found. The company says it had 20,000 users in October 2025 and more than 100,000 by July 2026. The page also shows earlier companies that promised to turn people's ideas into products and failed.
How much revenue does Buildpad make?
Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.
| Fact | Value | Grade | How we know | Sources |
|---|---|---|---|---|
| Monthly recurring revenue, 1 July 2025 | US$6,000 | C | Felix Heikka's Show HN post: "we just hit $6k MRR after 9 months since launching". He counts from the Product Hunt launch in October 2024. | [1] |
| Run rate from the July 2025 figure | US$72,000 a year | calc | Calculated: US$6,000 times 12. This is the headline value used for sorting on the site. | [1] |
| Total revenue, 17 June 2025 | US$33,000 | C | Founder post, eight months after the first paying customer. He linked a Stripe screenshot on imgur, which was not opened for this page. | [3] |
| Monthly recurring revenue, 9 April 2025 | US$3,600 | C | Founder post title "Got my SaaS to $3,6k MRR". He linked an MRR screenshot on imgur, not opened. In the same thread he wrote that they use Stripe for subscriptions. | [2] |
| Monthly recurring revenue, 25 February 2025 | US$2,700, total revenue about US$9,000 | C | Founder post. US$1,000 MRR came 54 days after launch and US$2,000 MRR 98 days after launch. No money spent on marketing up to then; paid ads had started the week before. | [4] |
| Monthly recurring revenue, 11 February 2025 | US$2,600, total revenue more than US$8,000 | C | Founder post, the first to name Buildpad and his brother. 4,000+ users, 1,000+ monthly active users. Traffic sources in his order: Reddit, direct, Google, X, Product Hunt. | [5] |
| Monthly recurring revenue, 25 January 2025 | US$2,300 from more than 100 paying customers | C | Founder post. It does not name the product, but its launch details (Product Hunt #4, 500+ upvotes, 475 users in 24 hours) are the same as on the company's own history page. | [6, 9] |
| Revenue per paying customer, January 2025 | at most about US$23 a month | calc | Calculated: US$2,300 divided by "100+" paying customers. | [6] |
| Paying customers, 16 December 2024 | 70 active paying customers, 2,400 users | C | Founder's first Show HN post. | [7] |
| Revenue, 26 November 2024 | US$2,200 in less than two months | C | Founder post that does not name the product. When a commenter pointed to a lower figure in his brother's posts, he answered that US$2,200 was total revenue and US$1,000 was MRR. | [8, 6] |
| Growth of MRR, January to July 2025 | 2.6 times | calc | Calculated: US$6,000 divided by US$2,300. | [1, 6] |
| Users | 20,000 in October 2025; more than 100,000 by July 2026 | C | Company history page and the rebrand post. These are sign-ups, not paying customers. | [9, 10] |
| Team | 2 people, the two brothers | C | "It's just my brother and I working hard on this project, doing marketing, building, customer support, everything ourselves" (July 2025). The company history says that for almost two years it was just the two of them and that it is building a team in 2026. | [1, 9] |
| Outside funding | none found | C | No source mentions investors. In May 2024 Felix was living off his savings. | [9] |
| Company | BUILDPAD S.R.L., registered in Romania | C | Privacy policy and terms of service. Hosting is in Germany. | [15, 16] |
| Time from first release to US$6,000 MRR | about 10 months | calc | Calculated: 19 August 2024 to 1 July 2025 is 10.4 months. From the Product Hunt launch in October 2024 it is 9 months, as the founder counts it. | [1, 11] |
How to read this table: a grade C value is a statement by the founders, in Hacker News posts or on the company's own pages. A "calc" value was computed on this page from the rows above it. No value is grade A or B.
| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Free trial | 0 | limited daily credits | |
| Pro | 39 | 200 credits a month, email support | |
| Max | 85 | 500 credits a month, premium support | |
| 100 extra credits | 20 one-time | Pro and Max only | |
| 500 extra credits | 100 one-time | Pro and Max only |
The pricing page shows only monthly prices and no yearly option. Credits are used up by AI actions: a simple message costs a fraction of a credit, research costs more. Unused credits roll over for one month. Trial users get a small daily credit allowance. Extra credits are sold to Pro and Max subscribers only. The founder wrote in July 2025 that the problem-research part was free.
When did Buildpad launch, and how did it grow?
Two failed products, then a product about not building failed products. Each entry carries its grade.
- July 2023CDavid Heikka, who had been the marketing and sales co-founder of a software business that reached US$30,000 in a month, starts a coding course. [11]
- March 2024CFelix Heikka leaves the Swedish army after five and a half years and moves to Budapest to start a business with his brother. [9]
- April to June 2024CTheir first product, AIConvert, an AI lead-qualification form, gets almost no sign-ups after paid ads, SEO and cold messages; more than US$800 goes to Facebook ads. A second idea, TinderRoast, also fails. [9, 11]
- July 2024CThe idea: an AI that remembers a project between chats and guides the founder through phases. They test it with a survey posted in founder subreddits, offering feedback on respondents' projects in return. [9, 6]
- 19 August 2024CFirst version released, built in about 30 days. 100 users in two weeks against a goal of 20. [11, 4]
- October 2024CProduct Hunt launch: #4 of the day with 500+ upvotes, 475 new users in 24 hours, 1,000 users after a week, and the first paying customers. [9, 5]
- 26 November 2024CFelix reports US$2,200 in revenue in less than two months. [8]
- 16 December 2024CFirst Show HN: 2,400 users and 70 active paying customers. [7]
- February 2025CUS$2,600, then US$2,700 MRR and about US$9,000 in total revenue. A free validation search of internet discussions is released. Paid advertising starts. [5, 4]
- 9 April 2025CUS$3,600 MRR. Hacker News commenters say the brothers keep posting similar stories. [2]
- 17 June 2025CUS$33,000 in total revenue eight months after the first paying customer. [3]
- 1 July 2025CSecond Show HN: US$6,000 MRR, research on Reddit and X, a canvas and new phases. Still only the two brothers. [1]
- October 2025C20,000 users. The product is renamed after the domain aicofounder.com, bought to win search traffic for the term. [9, 10]
- July 2026CAfter nine months the name goes back to Buildpad; the founders call the rename a mistake. More than 100,000 users. [10, 17]
- October 2026CLive, with Pro at US$39 and Max at US$85 a month, research agents, AI interviews and a website builder. [13, 14, 12]
Why did Buildpad work?
Four causes, in the order they mattered. Each one is a reusable lesson.
Test demand before building
After two failed products, they asked their target users first: a survey in founder subreddits about how people solved the problem, how much it hurt and what they would pay. They had to post it several times in different subreddits and offered feedback on the respondents' own projects in return. When the first version was ready they sent it to the same people, which gave the first users.
The buyers are the readers of build-in-public posts
The customers are early founders, and early founders read the Build in Public community on X, founder subreddits and Hacker News. Felix posted there daily, with revenue numbers in the titles. The product's own story, failing until they validated an idea, is also its sales argument. In February 2025 he listed Reddit as the first traffic source.
One strong launch day
The Product Hunt launch reached #4 of the day and brought 475 users in 24 hours, the newsletter feature, and the first paying customers on the same evening. US$1,000 MRR followed 54 days after launch, by the founder's count.
Free research, paid credits
The research that tells a founder whether a problem is real was free in 2025 and brings people in. Longer work, such as planning, many research agents and websites, costs credits, so the price follows the AI usage that the company pays its model providers for.
How it was built
These details come from the founders' posts and the company's own pages, grade C, except the page-source details, which anyone can see. [5, 4, 10, 15, 12, 14, 2]
- David wrote the code and Felix learned marketing; later both worked on product and marketing. The first version took about 30 days. In the earliest code the product was called "biz-builder".
- The AI comes from "US-based AI providers", large language model providers that the privacy policy does not name, plus a voice provider for speech-to-text and text-to-speech in AI-run interviews. No source names the models.
- Research agents work in parallel and read Reddit, X, app stores, Trustpilot, Kickstarter, market reports and other web pages, and return reports with links to their sources. No source says how the app gets access to Reddit and X posts.
- Hosting is with Hetzner in Germany; email goes through Mailgun. The page source of the site shows a Next.js build and Plausible analytics.
- Subscriptions are taken with Stripe, according to the founder. Advertising is measured with the Meta Pixel and Meta's Conversions API.
- Users publish the websites they build on buildpad.site subdomains or their own domains.
Tools named in the sources
Each one opens its page in the Tools tab, with every story that uses it.
Who tried this before Buildpad, and why did they fail?
Four companies in the Loot Drop database promised to help people turn their ideas into products, or to get quick answers from other people about them. The values below are from that database and are not independently verified.
Quirky
The promiseAnyone could submit an invention idea; the community voted, and Quirky built and sold the winners.
Why it diedIt made too many products without checking demand for each one, and spent money faster than sales came in.
SellanApp
The promisePeople with app ideas were matched with developers and backers who shared the revenue.
Why it diedA three-sided marketplace was too complex, and simpler, cheaper do-it-yourself app builders took the users.
Thumb
The promiseReal-time opinions from a crowd of other users on any question.
Why it diedIt never found a way to make money and was lost next to the large social networks.
Thinkature
The promiseA shared online whiteboard for brainstorming and planning projects in real time.
Why it diedGoogle Docs gave collaboration away free, and free users did not become paying customers.
The failed startups come from Loot Drop, a free database of dead startups, retold here in our own words. Loot Drop describes its entries as AI-assisted summaries of public sources, so these cards carry no evidence grade.
What changed
Quirky and SellanApp needed people, factories or developers to turn each idea into a product, and that cost money for every idea, good or bad. Thumb needed a crowd of human answerers and never charged them or the askers enough. Thinkature sold a shared canvas that a free product soon matched.
Buildpad replaces the crowd with AI agents that read the crowd's existing posts on Reddit and X, so an answer costs model usage, not people's time, and the user pays for that usage through credits. The canvas is not the product; the research and the guidance are. The free product that could match it is a general chatbot, and the company's answer is a guided process and research with cited sources that a plain chat does not give.
Can one person copy Buildpad now?
The method, yes. The figures, check them first.
Turning your own failed attempts into a product for people in the same position, and selling it where those people already read, is open to anyone. Founder tools are a crowded market, and many of them are sold with the same build-in-public posts, so the posts alone are no advantage.
Where it is still open: Research tools for one kind of buyer, for example people starting a shop, an app or a local service, that read public complaints and reviews for them and cite every source.
- 1Post a short survey where your target users meet, and give them something in return for answering.
- 2Send the first version to the people who answered the survey.
- 3Post progress with real numbers in the communities your buyers read, and launch once on Product Hunt when the product is ready.
- 4Give the first useful result away free and charge credits for the AI work that costs you money.
What could go wrong for Buildpad?
The risks are the mirror image of the causes above.
Every revenue figure comes from the founder
All the money figures are Felix Heikka's own posts on Hacker News between November 2024 and July 2025. They agree with each other and rise steadily, which makes them grade C, but the Stripe screenshots he linked on imgur were not opened for this page, Buildpad has no TrustMRR profile, and no journalist has seen the numbers. No revenue figure after July 2025 was found; the later numbers, 20,000 and 100,000 users, are sign-ups, not revenue.
It depends on model providers it does not name
The privacy policy says the AI comes from US-based large language model providers and a voice provider, without names. A price rise, a policy change or an outage at a provider would hit the cost of every credit, and the same providers sell general chatbots, such as ChatGPT, that the company has to argue against on its own comparison page.
It depends on reading Reddit and X
The feature that the product leads with reads posts on Reddit and X. No source says how Buildpad gets those posts. If either site limits or prices that access, the core research would get worse or more expensive.
Repeated promotional posts
On Hacker News, commenters complained in January and April 2025 that the brothers posted similar revenue stories many times; one counted 8 posts in sixty days. The posts were from the founders' own named accounts, not fake ones, but the same channel can turn against a product when readers see it as advertising.
Name changes and a growing team
The product was called aicofounder.com from October 2025 to July 2026 and then went back to Buildpad; the founders say the nine months could have gone into the Buildpad brand. The company history says it is building a team in 2026, so the two-person picture may not hold for later figures.
Compared with FounderPal
FounderPal, another two-founder tool on this site that turns a business description into a plan with an AI model, has grade A evidence for US$203K of total revenue in October 2026, while Buildpad's figures rest on the founder's word alone.
Sources
The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.
- C Hacker News, 1 July 2025: Show HN: I made an app to help you find problems with real demand from Reddit/X (Felix Heikka), via the Algolia APIRead in full through the Algolia API copy of the item, with the founder's reply. Source for the US$6,000 MRR, the team of two and the features in July 2025.
- C Hacker News, 9 April 2025: You're overcomplicating it. Just solve a real problem. (Got my SaaS to $3,6k MRR), via the Algolia APIRead in full through the Algolia API, with all comments. Source for the US$3,600 MRR, the use of Stripe and the commenters' criticism. The linked imgur screenshot was not opened.
- C Hacker News, 17 June 2025: Took 7 months to get my first customer. 8 mo to reach $33k revenue. Keep going., via the Algolia APIRead in full through the Algolia API, with the founder's replies (item 44300181 links a Stripe screenshot on imgur, not opened). Source for the total revenue steps up to US$33,000.
- C Hacker News, 25 February 2025: My app makes me $2,700/month after 6 months, via the Algolia APIRead in full through the Algolia API. Source for the US$2,700 MRR, the US$9,000 total, the 54 and 98 days to the first milestones and the start of paid ads.
- C Hacker News, 11 February 2025: My brother and I built a $2,600 MRR business in 6 months, via the Algolia APIRead in full through the Algolia API. Source for the US$2,600 MRR, the users, the traffic sources and the division of work between the brothers.
- C Hacker News, 25 January 2025: Growing my SaaS from 0 to 4k users in 5 months, what worked for me, via the Algolia APIRead in full through the Algolia API, with all comments. The post does not name the product; its launch details match the company history page. Source for the US$2,300 MRR, the paying customers and the survey method.
- C Hacker News, 16 December 2024: Show HN: Buildpad, Build products that people actually want, via the Algolia APIRead in full through the Algolia API. Source for the users and paying customers in December 2024 and the early features.
- C Hacker News, 26 November 2024: Validating my idea got me to $2200 revenue in less than 2 months, via the Algolia APIRead in full through the Algolia API. The post does not name the product; the founder confirmed the figure in the January 2025 thread. Source for the US$2,200.
- C Buildpad: Our story, read 26 October 2026Read in full. Company history with dates. Source for the founders' background, the failed products, the launch, the users in October 2025 and the team plan for 2026.
- C Buildpad blog, July 2026: aicofounder.com is now Buildpad (David Heikka)Read in full. Source for the rename in October 2025, the return to Buildpad, the 100,000 users and the earliest code name.
- C Buildpad blog: Learned to code, built a SaaS, now have paid customers from 40+ countries (David Heikka)Read in full. Source for David's earlier business, the coding course, the US$800 of Facebook ads and the 19 August 2024 release.
- C Buildpad blog, May 2026: Buildpad reviews: what it is and what founders are saying about it in 2026 (David Heikka)Read the sections on features and pricing. The company's own post. Source for the extra credit prices and the website builder.
- C Buildpad pricing page, read 26 October 2026Read in full. Source for the Pro and Max prices and credits.
- C Buildpad home page, read 26 October 2026Read in full, with the page source. Source for the research agents, the 100,000+ founders claim and the Next.js and Plausible details.
- C Buildpad privacy policy, last updated 29 August 2026Read in full. Source for the Romanian company, Hetzner, Mailgun, the unnamed US AI providers and the Meta Pixel.
- C Buildpad terms of service, read 26 October 2026Read in full. Source for the credit rules and the trial allowance.
- C Buildpad: What is an AI cofounder, 10 July 2026Read in full. Source for the research agents and the dates of the rename.
- C Loot Drop database export, 1,749 failed startupsLocal export. AI-written summaries, not independently verified. Source for the failed predecessors.