Skip to content
MakerGain
← Back to project

Half a Day to the First Version, a Payment Link on Day One: AudioPen Turns a Spoken Ramble Into Text You Can Send

Non-technical founder Louis Pereira used his own half-day build event to make AudioPen in about 12 hours: it transcribes scattered spoken notes and rewrites them in a chosen style into text that can be sent, or keeps the raw transcript. The app had a payment link on day one, and volunteers had already paid while trying it. Prime is sold in 3-month, 1-year, and 2-year terms; in May 2024 he said monthly revenue was about $15,000 and that he had been profitable from day one.

Speak into your phone for three minutes, changing your mind, repeating yourself, getting stuck, going wherever your thoughts go. AudioPen takes over exactly the recording nobody wants to listen back to: first it transcribes it as-is, then rewrites it in a style you choose into a piece of text you can send straight out. Rewriting is optional; you can also keep just the rough original transcript. The site's FAQ puts the difference between the phone's built-in dictation and AudioPen plainly—built-in dictation gives you back your exact words, including the “ums” and the half-finished sentences you start over; that is fine for a memo to yourself, but it looks bad in an email to someone else (FAQ[1]).

The person behind it is Louis Pereira, who describes himself as completely non-technical; by day he runs an offline business with his family, and at night he builds products online. In March 2023, as usual, he was trying the newest AI API of the moment and built 4 small tools in a week, all hosted on his personal website—he says this was to save the cost of buying domain names; he posted each to Twitter, and one got a noticeably better response than the others: AudioPen (Starter Story[2]). He has built more than 10 small products, most of them now dead; AudioPen is one of the few that survived (NoCodeLife interview[3]).

The 12-hour rule

The first version of AudioPen was built during the Half Day Build event he organized himself. By his account in Starter Story, the MVP took about 12 hours, followed by several months of improvements (Starter Story[2]). The event is held every other month, and its goal is stated bluntly: go from idea to revenue within 12 hours (NoCodeLife[3]). The rule forced him to cut down to the simplest version—“I was forced to build the simplest version people might pay for. I did. They paid.” (Starter Story[2])

The app had a simple payment link on it from day one. He says that because he was doing the Half Day Build at the time, it had to be that way (Starter Story[2]). Before the product officially launched, he found volunteers to try it early, and a group of them paid directly for the full version—he took that as a signal that “there is something here” (NoCodeLife[3]).

He built it with no-code tools—Figma for the interface, Bubble for the product, about $30 a month (Starter Story[2]).

Which step it picks up

The free version allows up to 3 minutes per note—enough for a thought jotted down by voice, not enough for an interview. Prime relaxes that to 15 minutes per note, and also lets you upload recordings already in your phone's recording app; they go through the same “transcribe and rewrite” path, and what comes back is finished text, not just a verbatim transcript (Pricing[4] · FAQ[1]). The strength of the rewrite can be adjusted across Low, Medium, and High. A style can also be described in a single sentence, or you can paste a piece of writing by yourself or someone else and have it imitate that; the free version can use all styles too, including ones you create (FAQ[1]).

The desktop version (Mac/Windows) changes the use case: hold a hotkey and speak, and the text is inserted directly at the cursor; only a double-press of the hotkey enters AI rewriting, and this rewrite path is processed in the cloud but not stored—discarded immediately after processing (AudioPen for Mac[5]). The Mac page lists very specific scenarios—replying to Gmail, writing Slack, drafting Google Docs, filling Jira tickets; the FAQ puts it more plainly: anywhere there is an input box (Mac[5] · FAQ[1]). There is no pile of “workflows” to set up; it writes wherever the cursor is.

Image: source AudioPen · official media kit[6], retrieved 2026-11-07; original interface preserved.

One thing it clearly does not do

The site's FAQ states the positioning crisply: AudioPen is for individuals thinking things through, not a meeting tool, so it does not join Zoom calls or label speakers. Most speech-to-text tools stop at transcription; AudioPen takes one optional extra step of rewriting (FAQ[1]).

Its boundary is drawn here too. Wispr Flow lists meeting notes as a separate tier, with Growth with Notetaker above that (Flow plans[7]). System-level dictation and AudioPen's desktop cursor voice input do overlap, but AudioPen also has style rewriting and a notes library, and it does not join meetings to label who said what (FAQ[1]).

Which step goes to the cloud, and where the cost falls

The privacy choice is not “whether to keep things secret” but which step secrecy lands on. The site says the audio of voice notes is deleted from its servers within 48 hours, the audio of voice input is never stored, data is encrypted at rest and is not used to train AI; iPhone, Android, Mac, and Windows can all turn on offline transcription, and once it is on, the audio does not leave the device (FAQ[1] · Mac[5]). But offline only covers transcription—the AI rewriting step itself runs in the cloud, and using it requires an internet connection. If you want the audio not to leave home, you can; if you want even the rewriting to be entirely local, the company does not offer that option.

Once notes are stored in the library, Prime users can let compatible AI assistants (such as ChatGPT and Claude) look through their notes and raw transcripts—read-only, with no ability to edit or delete (FAQ[1]).

How the money is charged

On the pricing page, Prime is sold by full term: 3 months for $33, 1 year for $99, 2 years for $159, and the page also shows the equivalent monthly prices of $11, $8.25, and $6.63; the page uses only the $ symbol and does not state the settlement currency (Pricing[4], retrieved 2026-11-14).

Image: source AudioPen · official pricing[4], retrieved 2026-11-07; original interface preserved.

There is only one dated revenue figure from an interview. In the 2024-05-16 Starter Story interview, he said monthly revenue at the time was about $15,000, that the business had been profitable from day one, that it was a one-person team, that the biggest cost was the AI API, that Bubble was about $30/month, and that he did not run paid ads or paid marketing; customer acquisition had come mainly from word of mouth from day one, with a referral commission program to encourage word of mouth and update emails to keep existing users engaged (Starter Story[2]). In the NoCodeLife interview, what he talked about was not money: less than 2 months after launch, there were already more than 1,000 paying users; his approach was to share on Twitter what he was building, ask for other people's opinions, and take them (NoCodeLife[3]). As for the first batch of testers, that is another thing he described in Starter Story: throughout the build he posted live on Twitter—in his words, live tweeted—which helped him build pre-launch buzz and brought the first beta testers (Starter Story[2]).

Of that, the $15,000/month was his own account in the 2024 interview and cannot represent current revenue. The only public costs are the AI API and the $30 for Bubble; refunds, customer support, and his own labor are not included, so profit remains unclear; there is likewise no data on renewal rates or retention.

When asked what exactly separates AudioPen from the products that did not take off, Louis's answer was: “I have absolutely no idea. I think it just happened to become a solution to a problem people did not realize they had.” (NoCodeLife[3])

Continue reading

AI productivity
Klipy, the sales AI that waits for your approval

Klipy takes over follow-up after sales meetings. It captures customer conversations from email, meetings, and instant messaging, drafts follow-ups and CRM updates, and prepares for the next meeting. Follow-up messages go out only after the salesperson approves them. Founder Jung Hong Kim is a Korean serial entrepreneur who took on debt when COVID sank his retail venture. He met his cofounder while consulting and launched Klipy.ai in three months. It went live in November 2024 as an AI CRM that logged communications automatically. In June 2025, it shifted toward carrying out next steps for sales teams and tied billing to specific actions users approve.

AI productivity
The founder reported \$8,000 in 2024 revenue, 85% collected only once: NodeLand turns long reports into maps

Rio de Janeiro-based Gabriel Xará runs NodeLand as a side project. Canvas extracts material from long reports, video transcripts, or meeting notes and generates structured maps; people still decide which nodes to keep, whether relationships are correct, and who the map is for. Built in 2023 and launched in 2024, it acquired early users through platforms such as AppSumo. A January 2025 interview disclosed \$8,000 in 2024 revenue, 85% from lifetime licenses, with costs and profitability undisclosed. He had focused almost entirely on the product that year, with limited marketing.

AI productivity
After closing a product nobody wanted, they broke entrepreneurship into “the next step”

David Heikka and his brother Felix launched their first AI form product and found zero demand. They tried ads, landing-page changes, renaming, and added features, then closed it after six months. They went on to build Buildpad, breaking business creation into stages that indicate what to validate next. Its early and current billing mechanisms differ, and the product was renamed aicofounder before changing back to Buildpad.

AI productivity
The client owns the code and pays the AI bill: SpaceAI360 builds automation you can take with you

Founded by Shahdin Salman alone, SpaceAI360 builds portable AI workflows inside customers' own accounts. It handles parsing, OCR, field extraction, database checks, and ERP/CRM writes after a form, email, or file arrives, routing uncertain items to human review. Rather than generic AI consulting, he starts with the highest-friction bottlenecks, assigning deterministic steps to n8n and judgment steps to Claude.