SoundBoost.ai
SoundBoost.ai is a web and phone service for audio mastering, the last processing step before a song is released: it sets the loudness, the balance between low and high frequencies, and the dynamics (the difference between the quiet and the loud parts) of the finished mix. The user uploads a song and writes what the result should sound like, or pastes a Spotify link as a reference. The founder described the method in March 2024: an LLM, a large language model, the kind of AI program that reads and writes text, interprets the request and sets a signal chain, a row of audio processors (saturation, compression, stereo width, limiting and others) that then changes the sound. The company says that it wrote this mastering engine itself and that the engine first measures genre, tempo and LUFS, loudness units relative to full scale, the standard measure of perceived loudness that streaming services use to play all songs at a similar volume. A second tool does stem splitting: it separates a finished song into its parts, called stems (vocals, drums, bass and other instruments). The founder's posts show that this tool runs HTDemucs, an open-source model published by a researcher at Meta, on graphics processors.
Berkan Cesur, a musician and former product manager in Istanbul, left his job in late 2022 and built the product with three friends. An archived copy of the company's About page from June 2023 says that the company was "founded by four friends". They released it on 8 February 2024 under the name Diktatorial Suite, earned US$450 in the first month by his count, and renamed it SoundBoost in June 2025 because, he wrote, 90% of the customers were American and could not pronounce the old name. An iPhone app followed on 15 August 2025, and an Android app is on Google Play today. TrustMRR, a public database of startup revenue, read the company's accounts at Stripe, the processor for card payments, which the website uses, and at RevenueCat, a service that records the subscriptions sold inside phone apps through Apple and Google. On 13 October 2026 it showed US$217,705 of revenue in the last 12 months, US$43,309 in the last 30 days and US$288,689 since the start. Monthly revenue stayed between US$5,200 and US$6,719 from November 2025 to February 2026 and then rose to US$42,678 in July 2026; 49% of the 12-month revenue came in the last three months. MRR, monthly recurring revenue, the sum of all active subscriptions with each one converted to its price per month, was US$25,874 from 2,675 active subscriptions.
The revenue, MRR and subscription figures are grade A: records of two payment providers, read through API keys, not a founder statement. They are revenue, not profit, and the part from RevenueCat is gross revenue: by RevenueCat's definition, what customers paid minus refunds, before Apple or Google keeps its commission. Two facts on this page differ from the short version of the story. This is not a one-person business: the About page named four co-owners in 2024 and names seven people and one adviser in October 2026, and a friend of the founder says in public that he invested. And the fast growth is recent: before October 2025 the company had collected about US$71,000 in 20 months. No source states what caused the rise in 2026; the page lists the changes that came before it, and what happened to earlier companies that sold tools to independent musicians.
The numbers, graded
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 |
|---|---|---|---|---|
| Verified revenue, last 12 months | US$217,705 | A | TrustMRR profile, revenue last synced on 13 October 2026 at 23:06 UTC, from Stripe and RevenueCat together. The headline figure of this page. The profile does not publish the split between the two providers. | [1] |
| Verified revenue, last 30 days | US$43,309 | A | Same profile. Last 7 days: US$12,010. The profile gives the growth of the 30-day figure as +51.2%. | [1] |
| Verified revenue, last 3 and last 6 months | US$106,752 and US$176,441 | A | Same profile. | [1] |
| Verified revenue, all time | US$288,689 | A | Same profile. The profile field for the founding date says 8 November 2023; the public release was on 8 February 2024. | [1, 8] |
| Verified revenue by calendar month | Nov 2025 US$5,948; Dec US$5,200; Jan 2026 US$5,959; Feb US$6,719; Mar US$9,146; Apr US$10,911; May US$14,425; Jun US$25,557; Jul US$42,678; Aug US$30,862; Sep US$35,978; 1 to 13 Oct US$20,728 | A | Same profile; it publishes no earlier months. These months add up to US$214,111; the remaining US$3,594 of the 12-month figure falls in the second half of October 2025 (calculated). | [1] |
| Share of the 12-month revenue that came in the last 3 months | 49% | calc | Calculated: 106,752 divided by 217,705. The last 6 months hold 81%. | [1] |
| Rise of monthly revenue from January to July 2026 | 7.2 times | calc | Calculated: 42,678 divided by 5,959. August was 27.7% below July, and September was 16.6% above August. | [1] |
| Revenue before the last 12 months | about US$70,984 | calc | Calculated: all time minus last 12 months, 288,689 minus 217,705. This covers about 20 months, from the release on 8 February 2024 to the middle of October 2025, an average of about US$3,500 a month. | [1] |
| MRR, monthly recurring revenue | US$25,874 | A | Same profile. Growth in 30 days: +52.6%. | [1] |
| Active subscriptions | 2,675 | A | Same profile. | [1] |
| MRR per active subscription | US$9.67 | calc | Calculated: 25,874 divided by 2,675. This is above the web yearly plan (US$4.00 a month) and below the web monthly plan (US$16.00), so the customers hold a mix of plans. | [1, 12] |
| Yearly rates computed from the two figures | US$519,708 from the last 30 days; US$310,488 from MRR | calc | Calculated: 43,309 times 12, and 25,874 times 12. Revenue is higher than MRR because a yearly plan counts in full in the month of sale, while MRR counts one twelfth of it. A month with many new yearly plans therefore shows more revenue than will repeat. | [1] |
| Founder's own revenue figures in 2024 | US$450 in the first month; US$1,516 in total by 19 April; US$2,000 by 5 May; US$3,000 by 28 May; MRR above US$3,500 at the end of July | C | Founder's posts. On 29 July 2024 he wrote that he would stop publishing revenue. These figures agree with the calculated average of about US$3,500 a month before October 2025. | [5, 8, 2] |
| Paying subscribers on 12 November 2025 | 1,000, with 33,000 users in total | C | Founder's post. TrustMRR's verified revenue for November 2025 is US$5,948, which is US$5.95 per paying subscriber (calculated). | [3, 1] |
| People named on the company's About page | 4 co-owners in May 2024; 5 people and 1 adviser in June 2025; 8 people and 1 adviser in January 2026; 7 people and 1 adviser in October 2026 | C | Live page and archived copies. The roles in October 2026: chief executive and product, design, front-end, design engineer, software developer, social media, customer support, and one adviser. The January 2026 list included an intern. The page does not say who works full time. On 3 May 2026 the founder wrote that his companies pay their interns. | [10, 11, 8] |
| Verified revenue per person on the About page | about US$31,100 a year | calc | Calculated: 217,705 divided by the 7 people listed in October 2026, without the adviser. This is revenue before every cost, not pay. | [1, 10] |
| Outside funding | one investor, amount not published | C | Batuhan İçöz wrote "I know I invested in the right startup" on 16 August 2024 and "i'm an investor" on 1 July 2025; the company's account answered a third such post on 20 September 2026 with "This guy believed us on day zero". The founder wrote on 6 February 2025 that this friend has invested in almost everything he has done. Yet on 16 January 2025 the founder called the company "a fully bootstrapped startup with no VC or investor backing". The statements do not agree. No venture round was found. | [9, 27, 8] |
| Users, by the company's own counts | 33,000 on 12 November 2025; 100,000 on 10 June 2026; 110,000 on 1 July 2026; 217,622 on the website on 13 October 2026 | C | Founder posts and the home page. The TrustMRR profile field "Estimated user count" says 628,739, which is close to the website's count of masterings (657,140), not to its count of musicians. | [3, 8, 7, 12, 1] |
| Share of users who make their songs with Suno | 60% | C | Founder's reply to a Suno employee on 1 July 2026, in which he asked for access to Suno's API. | [7] |
| Share of customers in the USA | 90% to 95% | C | Founder's posts of 5 June 2025 (90%), 4 December 2025 (95%) and 7 June 2026 (90%). He describes the company, DIKTATORIAL Inc., as American; he lives in Turkey. | [8, 12] |
| US App Store rating | 4.6 of 5 from 140 ratings | B | Read on 13 October 2026. Apple's data gives 4.64. Released on 15 August 2025; version 3.00 is dated 5 October 2026. The seller is DIKTATORIAL INC. | [13] |
| Google Play listing | 4.5 stars from 217 reviews, more than 10,000 downloads | B | Read on 13 October 2026; last updated on 30 September 2026. The release date of the Android app was not found. | [14] |
| Product Hunt launch, 26 March 2024 | 270 points | C | Product Hunt page read in October 2026. The company's About page says it finished fourth among the products of that day; the rank could not be read on the Product Hunt page. | [16, 10] |
| Marketing spend in the first two months | US$0 | C | Founder's posts of 15 March and 17 April 2024. The TrustMRR profile names word of mouth as the only marketing channel. | [5, 8, 1] |
| Landing page test, June and July 2024 | sign-ups rose from about 9% of visitors to about 30% | C | Founder's post of 29 July 2024, after two months of measuring two versions of the home page. | [2] |
| How the app revenue is counted | before the stores' commission | C | RevenueCat's documentation defines Revenue as the total paid minus refunds, and defines a separate figure, Proceeds, after the stores' taxes and commission. TrustMRR's FAQ says that it imports revenue from RevenueCat and does not mention a deduction. The share of app revenue inside the total is not published. | [17, 18] |
| Store commission for small developers | 15% at Apple; 15% or less at Google for 99% of paying developers | C | Apple's Small Business Program page, for developers with up to US$1 million of proceeds a year, and Google Play's service fee page. On a weekly plan of US$9.99 a 15% commission is US$1.50 (calculated). | [19, 20, 13] |
| Price of unlimited mastering at a music distributor | US$99 a year | C | DistroKid's Mixea page: the first track is free, then US$99 a year for unlimited tracks. SoundBoost's web yearly plan costs US$48. | [23, 12] |
| Listed for sale | no | C | The TrustMRR profile says that the startup is not listed for sale. Read on 13 October 2026. | [1] |
| Time from the release to the October 2026 figure | 32 months | calc | Calculated: February 2024 to October 2026. | [8, 1] |
How to read this table: a grade A value was read by TrustMRR from the company's Stripe and RevenueCat accounts through API keys; it is not a founder statement. A grade B value is published by Apple or Google in a store listing. A grade C value is a founder or investor statement, a page of the company, a profile field that the founder filled in, or a vendor's document. A "calc" value was computed on this page from the rows named in its note.
| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Unlimited, web | 4.00 | 16.00 | billed US$48 a year; 15 revisions per track |
| Unlimited Plus, web | 6.00 | 24.00 | billed US$72 a year; bulk mastering, unlimited revisions, drum, piano and guitar stems, priority queue |
| Unlimited, iPhone app | 4.17 | 17.99 | US$49.99 a year, or US$9.99 a week |
| Unlimited Plus, iPhone app | 5.92 | 24.99 | US$71.00 a year, or US$14.99 a week |
Every plan includes unlimited mastering and stem splitting; a free preview needs no account. On the web the yearly plans are billed at US$48 and US$72, which the page data shows as 75% below twelve monthly payments (US$192 and US$288). In the iPhone app the yearly plans cost US$49.99 and US$71.00, and there are weekly plans at US$9.99 and US$14.99, which do not exist on the web. A weekly plan kept for a full year costs US$519.48, about 10.8 times the web yearly plan (calculated). The first column is the yearly price divided by 12, computed on this page. The page data also holds two older plans with mastering credits, Pro at US$4.99 a month or US$34.99 a year, which the page does not show. Google Play prices were not read.
The timeline
From four friends who worked a year without income to US$43,309 of verified revenue in 30 days, 32 months after the release. Each entry carries its grade.
- Late 2022CCesur leaves a salaried job to work on music. He writes later that he had worked as a product manager for 12 years and has built websites for 20 years. He released his first own track in June 2020. [8, 3]
- 16 January 2023CHe announces Diktatorial Suite, an AI mastering service, for the first quarter of 2023. The release comes a year later. [8]
- 30 June 2023CSign-up for a closed test opens at diktatorial.com. The About page says the company was founded by four friends. [8, 11]
- 8 February 2024CPublic release of Diktatorial Suite. In the first four days: 68 new members and 18 full masterings. [8]
- 15 March 2024CFirst month: 263 musicians, 1,048 masterings, 20 paid subscriptions, US$450 of revenue, US$0 spent on marketing. Cesur moves to a village near Bursa to work on the product for at least a year. [5]
- 26 March 2024CLaunch on Product Hunt. He explains that an LLM interprets the request and controls the audio processors, and writes that four people built the product for a year without expecting income. [4, 16]
- 28 May 2024C5,000 masterings, 1,700 users and US$3,000 of revenue in total since the release. [8]
- July 2024CMusic reviewers on YouTube publish videos about the product; he writes that there is no commercial tie to them. A new home page triples the sign-up rate. MRR passes US$3,500, and he stops publishing revenue. [8, 2]
- 16 August 2024CHe writes that the older competitor eMastered runs ads with Diktatorial's name in the title, and offers 50% off to its customers who switch. His investor answers in public the same day. [8, 9]
- 20 August 2024CMixmag Türkiye writes about the product and describes a founding team of four with an adviser from Bilkent University. [15]
- October and November 2024CVersion 2.0 on 7 October. 30,000 masterings by 22 October, one every 15 seconds. A mastering API for selected partners on 8 November. A free Loudness Penalty tool that runs in the browser on 15 November. [8]
- 16 January 2025CFirst-year report: more than 15,000 users and 50,000 masterings, 20,000 of them from text requests. [8]
- 5 June 2025CThe product is renamed SoundBoost, with 26,000 users. On 17 June the third version of the mastering engine adds six AI mastering engineers with different styles. [8]
- 15 August 2025CThe iPhone app is released. Cesur wrote it himself in about four months with AI coding tools. Subscriptions in the app go through RevenueCat. [6, 13]
- 12 November 2025C1,000 paying subscribers and 33,000 users, by the founder's count. He writes that something changed for the product in the last few months. [3]
- November 2025AThe first month in TrustMRR's public history: US$5,948 of verified revenue. [1]
- February to July 2026AVerified monthly revenue rises from US$6,719 to US$42,678. [1]
- April 2026CThe product's X account is suspended for about two weeks and then restored. The founder writes that the team had uploaded a few videos through a tool called Post-Bridge and does not know the reason. [8]
- June 2026CThe servers move from Railway to a dedicated machine at Hetzner in Germany on 7 June. On 10 June: 100,000 musicians, 360,000 masters and 117,000 text requests. [8]
- 1 July 2026CCesur asks Suno for access to its API and writes that 60% of his 110,000 users make their songs with Suno. [7]
- August 2026CThe investor buys the domain soundboost.com and writes that the founder now owes him US$15,000. The company announces the new domain on 22 August. [9, 27]
- August and September 2026AVerified revenue is US$30,862 in August, 27.7% below July, and US$35,978 in September. [1]
- 16 September 2026CA Music Video Generator is released. Three days later lip sync is added, at no extra cost for subscribers during the launch period. [27]
- 13 October 2026ATrustMRR shows US$217,705 in the last 12 months, US$43,309 in the last 30 days, MRR of US$25,874 and 2,675 active subscriptions. [1]
Why it worked
Four causes, in the order they mattered. Each one is a reusable lesson.
Timing: AI song generators made new customers
The product was built for producers who mix their own songs. By July 2026 the founder counted 60% of users as people who make their songs with Suno. Such a user holds a finished song, has no mixing skills, and still needs the last step before a release. Customers' public posts list the tools in order: the song from Suno, the mastering from SoundBoost, the pictures and the video from other AI tools.
A written request, not a row of controls
The user writes "more punch, softer cymbals", and an LLM sets the processors. Each request returns a new revision that can be compared with the earlier ones. A person who does not know what a compressor does can still say what is wrong with the sound.
Unlimited use at a low price, weekly plans on the phone
On the web the yearly plan costs US$48 for unlimited masterings and stem splits. In the iPhone app, released in August 2025, the same service is also sold by the week at US$9.99. Verified revenue began to rise six months after the app appeared. The profile does not show how much of the revenue comes from the apps.
Musicians as the sales channel
The team spent nothing on marketing in the first two months. Reviewers on YouTube found the product in July 2024. Since 2025 customers post their songs on X and name the product, and the company shows customers' songs on its home page. The TrustMRR profile names word of mouth as the only channel. The founder also writes that posts on Reddit are a method he repeats.
Free tools that run in the browser
A loudness meter, a Loudness Penalty checker, a metadata editor, a vocal remover and a free mastering preview are open without payment. The first of them appeared in November 2024. No source measures how many subscribers they bring.
How it was built
These details come from the founder's posts, the company's pages and the TrustMRR profile fields. They are grade C. [4, 6, 8, 10, 12, 26, 1]
- Mastering: an LLM turns the written request into settings for audio processors: saturation, stereo width, compression, limiting, true-peak control and tape effects. The company says it wrote the engine itself, that a render takes about one minute, and that the current engine was trained on audio the company generated, not on customers' uploads.
- Stem splitting: HTDemucs, an open-source separation model, on graphics processors. In November 2025 the founder compared three graphics cards on a 3-minute track and wrote that the most expensive one was about 14% faster and that customers are happy with the speed.
- Web: Next.js for the pages, Python with Django for the server, Celery for the queue of audio jobs, Stripe for payments.
- Phone app: Expo and React Native, written by the founder in about four months with Cursor and Claude Code, his first app in ten years. RevenueCat handles the subscriptions and PostHog the usage statistics.
- Hosting: containers at Railway and Mac servers for audio processing in August 2025. After repeated instability he moved to a dedicated Hetzner server in Germany in June 2026 and wrote that response times fell to about a quarter, although 90% of customers are in the USA.
- Company: DIKTATORIAL Inc., which the founder describes as American. The people on the About page work in design, front-end and software development, social media and customer support.
Tools named in the sources
Each one opens its page in the Tools tab, with every story that uses it.
The dead that came before
Four companies in the Loot Drop database promised musicians or listeners a professional result without a studio, a label or expensive equipment. The values below are from that database and are not independently verified.
Jamglue
The promiseRemix and mix songs in the web browser, without studio equipment.
Why it diedProcessing audio in a browser was too hard at that time, the rights to the remixed songs were costly, and SoundCloud grew.
PonoMusic
The promiseStudio-quality sound for listeners, with its own player and download store.
Why it diedFew listeners paid for better sound quality while streaming on a phone was more convenient.
Earbits
The promiseOnline radio where independent artists pay to reach listeners without a label.
Why it diedPromotion paid by artists did not bring enough money, and the streaming services improved their own discovery of independent music.
Mint Songs
The promiseLet musicians sell their songs as NFTs and earn without the traditional industry.
Why it diedThe NFT market fell in 2022, and artists and fans lost interest.
What changed
Jamglue failed partly because a browser in 2006 could not process audio well. In 2026 SoundBoost renders a master in about one minute and runs its loudness meter inside the browser. PonoMusic and Mint Songs offered something that few musicians or listeners had asked for. Earbits took its money from independent artists, who have little of it, and the income was not enough.
Two things differ for SoundBoost. First, mastering is a step that a song needs before release, so the customer already has the task; the product only has to do it at a price the customer accepts, here US$4 to US$16 a month on the web. Second, AI song generators have multiplied the number of people who hold a finished song and cannot master it. The weak point is the one that Earbits had: the payers are independent musicians, the price is low, and larger companies can add the same step to what they already sell.
Can one person copy it now?
The pricing and the positioning, yes. The audio engine, only with the skills.
The team had a founder who could judge sound, a developer for signal processing, a year without income before the release, and then 20 months at about US$3,500 a month. That part is expensive to repeat. What can be copied is the shape: one finishing step that every maker of some kind of content must do before publishing, done from a written request, sold as an unlimited subscription on the web and by the week on the phone, next to a fast-growing generator that leaves the step undone.
Where it is still open: Finishing steps for other AI-made media: cleaning and levelling AI voice recordings for podcasts, sound and colour finishing for AI video, and print preparation for AI images.
- 1Find the finishing step that a popular AI generator leaves undone, and ask its users where they stop.
- 2Let the user ask in plain words, and return every request as a new version that can be compared with the last one.
- 3Put free single-purpose tools on the site that run in the visitor's browser, so a first visit costs you no computing.
- 4Sell the same subscription on the web and in the phone app, and test weekly plans in the app.
- 5Show customers' work on your own pages and accounts, so that they have a reason to name the product in public.
What to watch
The risks are the mirror image of the causes above.
Seven people, and revenue is not profit
The verified figures are revenue. Seven people and an adviser are listed on the About page, which gives about US$31,100 of revenue per person per year before any cost. No source publishes salaries, profit or the amount the investor put in, and the founder and the investor describe the funding differently.
Most of the revenue is three months old
49% of the 12-month revenue came in the last three months. July 2026 was the highest month, and August was 27.7% lower. The 30-day revenue (US$43,309) is well above MRR (US$25,874), so a large part is new sales and yearly plans paid in full, sold at 75% below the monthly price. Yearly plans sold since the rise began have not reached a renewal date yet.
The cost of audio processing with unlimited plans
Every master and every stem split uses computing time, and the founder runs stem splitting on graphics processors. He compares cards for speed, and in 2026 he moved to another host after instability that, he wrote, was losing him customers. The plans are unlimited from US$4 a month, so a heavy user costs more than a light one and pays the same. The company publishes no cost figures.
Older and larger competitors
LANDR says that more than 5 million musicians trust its AI mastering and also sells a plugin and an API. iZotope's Ozone has had an AI Master Assistant since 2017. eMastered, which describes itself as an online mastering engine made by Grammy-winning engineers and powered by AI, ran ads with the product's old name in 2024, the founder wrote. DistroKid, a music distributor, sells unlimited mastering under the name Mixea for US$99 a year, inside the service that musicians already use to release songs. BandLab's mastering page could not be opened for this research, so nothing about it is verified here.
Dependence on Suno's users
By the founder's count, 60% of users come with songs made in Suno, and he has asked Suno for API access. Suno's own pricing page already lists stem separation in its paid plans, from US$8 a month billed yearly, and in its highest plan a studio with effects and automation. If Suno adds a mastering step, the largest group of customers may no longer need a separate product.
App stores and platforms
Subscriptions sold in the apps pass through Apple and Google. Apple's Small Business Program sets the commission at 15% for developers below US$1 million of yearly proceeds, and Google's fee page says that 99% of paying developers qualify for 15% or less. The RevenueCat part of the verified revenue is counted before this commission. The stores decide whether an app stays listed, and the product's X account was suspended for two weeks in April 2026 without a stated reason.
Limits of the evidence
TrustMRR publishes monthly figures only from November 2025, does not split Stripe from RevenueCat, and says that its own figures can differ from a provider's dashboard by up to 30%. The founder stopped publishing revenue in July 2024. User counts are the company's own and rise very fast: from 33,000 to 217,622 in eleven months.
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.
- A TrustMRR profile of SoundBoost.ai (formerly Diktatorial), revenue last synced 13 October 2026Read the markdown profile in full on 13 October 2026. The revenue, MRR and subscription figures come from the company's Stripe and RevenueCat accounts through API keys and are grade A. The description, founding date, user count, tech stack and marketing channel are founder-written fields, grade C.
- C Berkan Cesur on X, 29 July 2024: MRR above US$3,500, and the result of the landing page testRead in full through api.fxtwitter.com, in Turkish, with the post it quotes. Source for the last revenue figure he published and for the sign-up rates.
- C Berkan Cesur on X, 12 November 2025: 1,000 paying subscribers and 33,000 usersRead in full through api.fxtwitter.com. Source for the subscriber count and for the date he left his job.
- C Berkan Cesur on X, 26 March 2024: how the LLM controls the signal processing, and four people working a year without incomeRead in full through api.fxtwitter.com, in Turkish. Source for the method and for the number of people who built the product.
- C Berkan Cesur on X, 15 March 2024: report on the first monthRead in full through api.fxtwitter.com, in Turkish. Source for the first-month counts, the US$450 and the zero marketing spend.
- C Berkan Cesur on X, 15 August 2025: the iPhone app and its tech stackRead the thread of five posts through api.fxtwitter.com. Source for the app release, the four months of work, RevenueCat, Django, Celery, Railway and the Mac servers.
- C Berkan Cesur on X, 1 July 2026: 60% of 110,000 users create with SunoRead in full through api.fxtwitter.com. A reply to a Suno employee.
- C Berkan Cesur's X timeline, September 2021 to October 2026Read 732 posts page by page through api.fxtwitter.com, many of them in Turkish. Source for the dated events not covered by the single posts above: the 2023 announcements, the release on 8 February 2024, the 2024 statistics, the rename, the customer shares by country, the server move, the user counts and his statements about funding and staff.
- C Batuhan İçöz on X, 1 July 2025: "@soundboostai (i'm an investor)"Read in full through api.fxtwitter.com, together with his posts of 16 August 2024, 14 August 2026 and 20 September 2026. Source for the investment and for the purchase of the .com domain. No amount is stated.
- C SoundBoost About page, read 13 October 2026Read in full, with the script file that holds the team list, because the list loads by script. Source for the company's own history, the engine statements and the eight names with their roles.
- C Archived copies of the About page, 2023 to 2026 (Wayback Machine)Read the copies of 30 June 2023, 29 May 2024, 17 January 2025 and 13 June 2025, and the team script of the copy of 29 January 2026. Source for the size of the team at each date.
- C SoundBoost pricing page and home page, read 13 October 2026Read both pages in full. The page shows the yearly price; the monthly prices and the older plans were read from the plan data inside the page source. Source for the web prices, the feature lists, the free tools, the company name and the website's own counts.
- B US App Store listing: Audio Mastering - SoundBoost, by DIKTATORIAL INCRead the listing and Apple's lookup data on 13 October 2026. Source for the release date, the version, the rating and the six in-app purchases.
- B Google Play listing: Audio Mastering - SoundBoostRead on 13 October 2026. Source for the stars, the number of reviews, the download range and the date of the last update.
- C Mixmag Türkiye, 20 August 2024: Diktatorial Suite, a new online mastering tool supported by AIRead in full, in Turkish. A short article that repeats the company's statement. Source for the founding team of four and the adviser.
- C Product Hunt: DIKTATORIAL Suite, launched 26 March 2024Read the page text. The launch date and the points were readable; the rank of the day was not.
- C RevenueCat documentation: Revenue chartRead the definitions. Revenue is the total minus refunds; a separate figure, Proceeds, subtracts the stores' taxes and commission. Also read the page on the MRR chart.
- C TrustMRR FAQRead the sections on verification and on RevenueCat. It says that TrustMRR imports revenue, MRR and subscription history from RevenueCat and that its figures can differ from a provider's dashboard by up to 30%. It does not say that store commission is subtracted.
- C Apple: App Store Small Business ProgramRead the section on the 15% commission and the limit of US$1 million of proceeds.
- C Google Play Console Help: Service feesRead the overview. Source for the statement that 99% of paying developers qualify for a fee of 15% or less.
- C LANDR: AI Mastering, Online Audio MasteringRead in full. Source for LANDR's own claim of more than 5 million musicians and for its plugin and API. Its prices load by script and were not read.
- C eMastered pricing pageRead the visible text; the prices load by script and were not read. Source only for how the company describes itself.
- C DistroKid: Mixea, instant online music masteringRead in full. Source for the price of US$99 a year for unlimited mastered tracks.
- C iZotope: Ozone 12Read the product description. Source for the AI Master Assistant and the year 2017.
- C Suno pricing pageRead in full on 13 October 2026. Source for the stem separation in the Pro and Premier plans and the studio in the Premier plan.
- C Demucs README on GitHub (facebookresearch/demucs)Read the README. Source for what HTDemucs is: a Hybrid Transformer model under the MIT license that separates drums, bass, vocals and other stems. The author writes that the repository is no longer maintained.
- C SoundBoost's X account, February 2024 to October 2026Read 227 posts page by page through api.fxtwitter.com. Source for the product announcements, the new domain, the Music Video Generator and the reply to the investor.
- C Loot Drop database export, 1,749 failed startupsLocal export. AI-written summaries, not independently verified. Source for the failed predecessors.