# AudiowaveAI

Aiolum story: https://aiolum.com/story/audiowaveai/ · added 2026-11-06 · updated 2026-11-06

**In short.** AudiowaveAI (AI text-to-speech that turns articles, books and PDFs into long-form listening playlists, Canada): About US$5K a month self-reported, November 2026. Evidence grade D: reported, not verified. Team: 1 founder, Michael Yagudaev in Vancouver; whatitmakes lists 1 founder. Launched May 2024; still running, revenue flat.

AI text-to-speech that turns articles, books and PDFs into long-form listening playlists · Canada · AI-native · solo founder · still running, revenue flat · group: Voice and audio

- Outcome: About US$5K a month self-reported, November 2026 (evidence grade of the headline number: D; headline in US dollars: 60000)
- Launch: May 2024; months from launch to the outcome: 30
- Team: 1 founder, Michael Yagudaev in Vancouver; whatitmakes lists 1 founder
- Outside funding: None reported
- Main factor: Price
- Meters, 1 to 4: evidence strength 1, copyable pattern 3, capital needed 1
- Tags: AI-native, Solo founder, Bootstrapped, Running, B2C, Text to speech

Links:

- Visit audiowaveai.com: https://www.audiowaveai.com/

**The product.** A web app, also usable as a mobile web app, that converts long text into spoken audio: articles, blog posts, EPUB books, PDFs or pasted text. It splits a book into chapters and long chapters into parts of about one hour, keeps the results in playlists, and gives each user an RSS feed so the audio plays in an ordinary podcast player. The home page lists 10 stock voices, and the API documentation says the voices come from OpenAI and Azure. The pricing section lists pay-as-you-go credit packs, not a subscription, and a beta API sells the same conversion to developers.

**Why it worked.** The founder wrote on Hacker News that he built it because a book he had bought, Make by Pieter Levels, had no audiobook version, and the text-to-speech options he tried were too robotic, too complex or too costly. His launch post states that only 3% of books exist as audio. He chose OpenAI's text-to-speech model after testing others and said that on Hacker News, where he also put the price of US$15 for about 10 hours of audio against a model cost of about US$10.

Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; E = no revenue figure found, so the headline is a non-money fact and the other facts carry their own grades; calc = computed on this page from graded facts.

## Summary

AudiowaveAI is a text-to-speech (TTS) service, a program that reads written text aloud with a synthetic voice. It is made for long-form listening: a user uploads an EPUB book or a PDF, or pastes a link to an article, and gets back hours of audio split into chapters, in a playlist that can also be fed to a podcast player through RSS, a standard web feed format that podcast apps subscribe to. The voices are not its own. The API documentation says "We support a wide range of voices from OpenAI and Azure", and the founder wrote on Hacker News in May 2024: "Just OpenAI TTS for now".

Michael Yagudaev, a software developer in Vancouver, Canada, introduced it on his blog on 7 May 2024 and launched it on Product Hunt and as a "Show HN" post on Hacker News on 23 May 2024. He wrote that it was project 12 in 16 weeks of a challenge to ship 52 products in 52 weeks. His launch report lists 1,158 visitors, 95 new users, 3 customers and US$45 of revenue on launch day. The product is still sold in November 2026, with a beta developer API and a footer dated 2026.

The headline figure is weak evidence. The revenue directory whatitmakes lists about US$5,000 a month, marked "Self-reported", flat from 6 October to 5 November 2026, and the directory Arrfounder lists the same US$5,000 a month and a founder card that reads "audiowaveai.com ＄5k". Both appear to rest on the founder's own public profile, and no payment processor confirms the figure, so the page grades it D.

## Pricing

AudiowaveAI prices from the pricing section of its home page, read 6 November 2026 (grade C)

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Free | 0 | 0 | 2 articles, up to 10 minutes of audio, 10,000 characters |
| 10 Hours of Audio | 15 one-time |  | 100 articles, up to 10 hours of audio, 600,000 characters; shown as 50% off US$30 |

The site sells credits, not a subscription: "No, AudiowaveAI is a pay-as-you-go service", and "No credits do not expire". The paid pack is shown as US$15 with US$30 struck through and the words "Limited time early access pricing 50% off". The FAQ offers a 30-day money-back guarantee. The FAQ says "Each word in the input text costs 1 credit" and then gives a 100-character example that costs 100 credits; the API documentation says each character costs one credit. A /pricing page returned 404, so the home page section is the only price list read.

## How much revenue does AudiowaveAI make? The numbers, graded

A solo text-to-speech product launched in May 2024, listed at about US$5,000 a month by two revenue directories in November 2026, both self-reported

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Revenue per month, 5 November 2026 | about US$5,000 | D | whatitmakes profile, "Self-reported", tracked since 25 September 2026; its Q&A says "We last checked on Nov 5, 2026" and "Not yet" verified. | [1] |
| Change in the last 30 days | 0% | D | whatitmakes: "Last 30 days 0% Growth in MRR", from US$5,000 on 6 October 2026 to US$5,000 on 5 November 2026. | [1] |
| Second listing of the figure | MRR US$5,000, ARR US$60,000 | D | Arrfounder's US$10K-100K ARR products page, downloaded 6 November 2026. Arrfounder says its data is "publicly disclosed by founders themselves and collected from various sources, primarily Twitter/X", so it is probably the same founder claim, not a second check. | [2] |
| Founder's own profile line | "audiowaveai.com ＄5k" | D | Founder card on Arrfounder's US$100K-1M ARR founders page, downloaded 6 November 2026. The line does not say whether US$5,000 is per month or in total. | [3] |
| Run rate from the November 2026 figure | about US$60,000 a year | calc | Calculated: US$5,000 times 12. whatitmakes shows the same "$60K annualized". | [1] |
| Packs needed for US$5,000 a month | about 333 packs of US$15 | calc | Calculated: 5,000 / 15. The site sells only one-time credit packs, so the figure implies about 333 pack sales a month, unless API use or larger purchases add to it. | [1, 4] |
| Launch-day results, 23 May 2024 | 1,158 visitors, 95 users, 3 customers, US$45 revenue | C | Founder's blog report on the Product Hunt launch: 138 upvotes, rank 14 of the day. | [7] |
| Hacker News launch | 55 points, 50 comments | C | "Show HN: Affordable text-to-speech for long-form content", 23 May 2024, read through the Hacker News search API. | [9] |
| Model cost and margin, May 2024 | about US$10 of OpenAI cost for about 10 hours of audio; margin about 50% at US$15 | C | Founder's reply on Hacker News: "Using OpenAI TTS, it costs ~$10 for ~10hrs. So at $15, margin is 50%." | [9] |
| Voice providers | OpenAI and Azure | C | API documentation, read 6 November 2026. In May 2024 the founder said it used only OpenAI TTS. | [5, 9] |
| Voices and languages | 10 voices; "Most voices support 100 different languages" | C | Home page FAQ. Further voices are added "upon request"; no page read offers to copy a real person's voice. | [4] |
| Stack | Next.js on Vercel with a Postgres database, Tailwind CSS | C | Founder's replies on Hacker News, May 2024. | [9] |
| Team | 1 founder | C | Blog posts are written in the first person and signed by Michael Yagudaev; the footer says "Made with ❤️ in Vancouver 🇨🇦 by @yagudaev"; whatitmakes lists "1 founder". | [1, 4, 6] |
| Outside funding | none reported | C | No source read mentions investors or a funding round. | [1, 4] |
| Time from launch to the November 2026 figure | 30 months | calc | Calculated: May 2024 to November 2026. | [1, 6] |

How to read this table: a grade D value is a figure from a revenue directory that marks it self-reported, with no payment-processor check. A grade C value is a statement by the founder on his blog, his site or Hacker News. A "calc" value was computed on this page from the rows above it. No value is grade A or B.

## When did AudiowaveAI launch, and how did it grow? Timeline

From a book with no audiobook to a self-reported US$5,000 a month, in 30 months. Each entry carries its grade.

- **Early 2024** (grade C [7, 9]): The founder runs a challenge to ship 52 products in 52 weeks. AudiowaveAI is project 12 in 16 weeks; QuickReadAI, a Chrome extension that summarizes text, launched a month before it.
- **7 May 2024** (grade C [6]): Blog post "Introducing AudiowaveAI: Text to Audio": he built it to turn a book that had no audiobook into audio.
- **23 May 2024** (grade C [7, 9]): Product Hunt launch, 138 upvotes, 95 new users, 3 customers and US$45 of revenue that day. The same day, a Show HN post gets 55 points and 50 comments.
- **30 August 2024** (grade C [8]): Update: an RSS feed per user so the audio plays in podcast players, processing of all chapters at once, and splitting of long chapters into parts of about one hour.
- **19 November 2024** (grade C [10]): The founder posts "All of Paul Graham's essays as Audio", a public playlist on the site, to Hacker News (3 points).
- **25 September 2026** (grade D [1]): whatitmakes starts tracking the product with a self-reported figure.
- **5 November 2026** (grade D [1, 2, 5]): whatitmakes lists about US$5,000 a month, unchanged over 30 days; Arrfounder lists the same figure. The site offers a beta API for developers.

## Why did AudiowaveAI work?

### A narrow use of a rented model

OpenAI released a text-to-speech model that sounds far less robotic than older system voices. AudiowaveAI wraps it for one job, hours of non-fiction listening, and adds the parts the raw model does not have: EPUB and PDF import, chapter splitting, a player, playlists and an RSS feed.

### Price set from the model cost

The founder wrote in May 2024 that 10 hours of audio cost him about US$10 at OpenAI, and he priced the pack at US$15. Credits do not expire and there is no subscription, which keeps the offer simple but means every month's revenue has to be sold again.

### Launches and content pages

The documented arrivals are a Product Hunt launch and a Show HN post. The site also hosts public playlists, such as Paul Graham's essays and guides on going to market, which are pages that can be found in search; no source says how many customers came from them.

## How it was built

These details come from the founder's blog, his Hacker News replies and the site's own pages. They are grade C.

- Next.js hosted on Vercel, a Postgres database and Tailwind CSS, by the founder's account on Hacker News in May 2024.
- Voices from OpenAI's text-to-speech model, and by November 2026 also from Azure, according to the API documentation. He said he tested other models and open-source options and found them worse in quality.
- Input text is converted to Markdown first; chapters are split on Markdown headings, and long chapters into parts of about one hour.
- A progressive web app (a web site that installs on a phone like an app) for listening, plus a per-user RSS feed for podcast players.
- A beta REST API with bearer-token keys, playlists and tracks, billed from the same credit balance.
- Built by one person as part of a series of fast launches; launch material was reused from his earlier product QuickReadAI.

Sources: [5, 7, 8, 9]

## Who tried this before AudiowaveAI, and why did they fail?

Four companies in our database of failed startups tried to turn text or information into listening or a richer reading format, and failed. The values below are from that database and are not independently verified.

The failed startups come from our database of dead startups, retold here in our own words; Its entries are AI-assisted summaries of public sources, so they carry no evidence grade.

### Voicery (2018 to 2021; funding US$2M; cause: Competition)

- Promise: Natural-sounding synthetic voices sold to developers through an API, for audiobooks, assistants and support bots.
- Why it died: Google, Amazon and Microsoft offered cheaper text-to-speech bundled with their clouds, and Voicery could not justify a premium price.
- Record: 

### Qwiki (2010 to 2015; funding US$10M; cause: No Market Need)

- Promise: Automatically narrated multimedia presentations instead of text search results.
- Why it died: High processing and storage costs and users who kept reading text; it was sold to Yahoo and closed.
- Record: 

### Vook (2009 to 2017; funding US$5M; cause: No Market Need)

- Promise: E-books with video and audio built into the text.
- Why it died: Readers kept reading plain e-books on Kindle, and publishers did not adopt the format.
- Record: 

### Post.fm (2010 to 2013; funding US$120K; cause: Competition)

- Promise: Personal radio stations built from audio clips collected from many sources.
- Why it died: Spotify's catalog and SoundCloud's community took the listeners, and it had no money to compete.
- Record: 

## What changed

Voicery tried to sell its own voice model and lost to the cloud companies' cheaper bundled voices. Qwiki paid for its own heavy processing. AudiowaveAI does neither: it rents the voice from OpenAI and Azure per character, so it has no model to train and its cost moves with each sale. The founder said in May 2024 that this left a margin of about 50%.

What did not change is the competition. The same cloud companies that beat Voicery sell their voices directly, phones and browsers read text aloud for free, and the founder himself named open-source models that improve every month. A small product like this one lives on the work around the voice: import, chapters, playlists and a feed for podcast players.

## Can one person copy AudiowaveAI now?

**Verdict: Yes, the method is open to one person; the figure is not checked.**

A solo developer built and launched it inside a weekly shipping challenge, on a rented model, with a standard web stack. Anyone can rent the same voices, so the copyable part is the choice of one narrow use and the work around it. The founder said in May 2024 that the model cost was two thirds of the price, which leaves little room for a lower price than his.

1. Pick one long-form job that a raw model API does badly: whole books, chapter splitting, a feed for an existing player app.
2. Price from the model cost per unit of output, and state the unit plainly (characters or hours, not both).
3. Launch on Product Hunt and Hacker News, and answer every technical question in the thread.
4. Add a developer API on the same credit balance once users ask for it.

**What is still open.** Long-form listening for one audience: course readings for one profession, company documents, newsletters in a second language.

## What could go wrong for AudiowaveAI?

- **What is not verified.** The headline of about US$5,000 a month rests on the founder alone. whatitmakes marks it "Self-reported" (tracked since 25 September 2026, last checked 5 November 2026), and Arrfounder says its figures are disclosed by founders and collected mainly from X; the founder's own profile line reads "audiowaveai.com ＄5k" and does not say per month. No payment-processor profile, interview or filing was found. At the US$15 pack price the figure would mean about 333 pack sales a month. A reader could look for a Stripe-verified profile or a dated revenue post by the founder.
- **Competition.** Among stories on this site, MicMonster also sells AI text-to-speech with stock voices, aimed at voice-overs rather than long-form listening. The founder's own Hacker News thread shows the other competitors: built-in system voices on Mac and iPhone, open-source models such as Piper, and the cloud providers' own APIs.
- **Depends on rented voices.** The voices come from OpenAI and Azure (API documentation). In May 2024 the founder said the OpenAI cost was about US$10 per 10 hours of audio against a US$15 price; a price rise or a policy change at the provider moves the margin directly.
- **Other people's texts.** The site hosts public audio playlists made from other authors' writing, such as Paul Graham's essays and guides by other founders. The FAQ says users must "comply with copyright laws for the original text if you are not the original author"; no source read says whether those authors gave permission.
- **Marketing wording and the founder's other products.** The home page has said "Limited time early access pricing 50% off" and "Limited to first 100! So act fast" on a product sold since 2024, and "Trusted by innovative companies and freelancers worldwide" without naming them. The FAQ and the API documentation define a credit differently (a word against a character). The footer links the founder's other products, including PHContacts, whose site did not answer on 6 November 2026, so what it sells was not checked. The blog's last post is from September 2024, an experiment with AI-generated guided meditation audio.

## Sources

1. whatitmakes: How much does Audiowaveai make? $5K/month revenue (2026) — https://whatitmakes.com/startup/audiowaveai (grade D). Read in full on 6 November 2026. Source for the self-reported US$5,000 a month, the 30-day change, the tracking start date and the founder count.
2. Arrfounder: $10K-100K ARR SaaS Products (2026) — https://arrfounder.com/products/10k-to-100k-arr (grade D). Read the AudiowaveAI entry in a copy downloaded 6 November 2026 UTC by the pipeline's harvest (research/harvest-search-api-20261106c/36); a new request on 6 November 2026 answered 429, so it was not fetched again. Source for the MRR and ARR listing and Arrfounder's data disclaimer.
3. Arrfounder: $100K-1M ARR Founders (2026), page 3 — https://arrfounder.com/founders/100k-to-1m-arr?page=3 (grade D). Read the founder card in a copy downloaded 6 November 2026 UTC by the pipeline's harvest (research/harvest-search-api-20261106d/11). Source for the line "audiowaveai.com ＄5k".
4. AudiowaveAI home page with pricing and FAQ — https://www.audiowaveai.com/ (grade C). Read in full on 6 November 2026. Source for the product, the prices, the credit rules, the voices and the footer.
5. AudiowaveAI API documentation (beta) — https://www.audiowaveai.com/docs/api (grade C). Read in full on 6 November 2026. Source for the OpenAI and Azure voices and the per-character credits.
6. AudiowaveAI blog, 7 May 2024: Introducing AudiowaveAI: Text to Audio — https://www.audiowaveai.com/blog/2024-05-07-intro (grade C). Read in full. Source for the origin and the launch month.
7. AudiowaveAI blog, 29 May 2024: Lessons from Launching on ProductHunt — https://www.audiowaveai.com/blog/2024-05-29-post-product-hunt-launch (grade C). Read in full. Source for the launch-day results and the 16-week challenge.
8. AudiowaveAI blog, 30 August 2024: July + August 2024 Update — https://www.audiowaveai.com/blog/2024-08-30-july-august-2024-updates (grade C). Read in full. Source for the RSS feed and the chapter features.
9. Hacker News, 23 May 2024: Show HN: Affordable text-to-speech for long-form content — https://news.ycombinator.com/item?id=40452411 (grade C). Read the whole thread through the Hacker News search API (hn.algolia.com/api/v1/items/40452411). Source for the founder's account of the model, the cost, the margin and the stack.
10. Hacker News, 19 November 2024: All of Paul Graham's essays as Audio — https://news.ycombinator.com/item?id=42185592 (grade C). Read the title, date and points in the Hacker News search API results; the thread itself was not opened.
11. AudiowaveAI blog index — https://www.audiowaveai.com/blog (grade C). Read in full on 6 November 2026. Source for the dates of the posts, the last being 19 September 2024.
12. Database of failed startups, local export — /method/#dead (grade C). Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Contact: aiolum@livyden.com. Run AudiowaveAI? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: AudiowaveAI).