All projects

WeDo.agency

A business consulting and development agency running many small, parallel client engagements rather than a single product — from AI-powered tools to e-commerce storefronts to internal ops dashboards.

Work

Audioland — AI Music Generation Plugin

WeDo.agency07/2023 – 10/2023
Shut down by OpenAI's plugin monetization policy, not by the product itself.
<1 min
Generation time

Audioland (and its sibling project, Dubbing) existed to take speechki.com public on the Cyprus stock exchange — the plan required demonstrating real profit from speechki.com's neural-network products. The startup didn't make it, not for lack of product, but because of OpenAI's plugin monetization policy: "plugins could not be monetized before, during, or after operation" — meaning not at all.

How a Cyprus IPO actually pays out

Taking speechki.com public meant more than good numbers — it meant walking this exact sequence, where only one party's payday is fixed and everyone else is betting on where the market lands.

  1. Show at least $50K in real, audited profit
  2. Spin the IP into a fresh holding entity, clean of the old investor cap table
  3. An independent audit sets the company's valuation and verifies its books
  4. A broker buys the shares at a discount to that valuation — this is the company's actual payday
  5. Shares start trading freely on the exchange — the company never sees another cent from that
Drag the slider below to set the sale price.
Price the company sold for$150,000

An independent audit estimated the company's value at $150,000

The market eventually valued the company at $150,000

Price upPrice downAudit estimateBroker's buy-in price
$0
Companyvs. market value
$150,000$150,000 = $0
Locked in the moment the broker buys — the market afterward doesn't change it
+$0
Broker
$150,000$150,000 = +$0
The market agreed with the audit — no windfall left on the table
$0
Exchange
Small trading fees every period, in either direction — the company sees none of it

Work

Dubbing — AI Video Translation & Dubbing

WeDo.agency05/2020 – 06/2021
Agency pivoted; product was client-owned and handed off.
150
First users
60
Languages
60×59
Programmatic landing pages

Programmatic SEO pages on Next.js targeting 60×59 language-pair search results brought the product's first 150 users at effectively zero acquisition cost — a scrappy alternative to paid marketing for an early-stage product with no budget. The go-to-market plan also included a listing on AppSumo — a marketplace selling one shared subscription across dozens of small SaaS tools — competing for the same customers as rask.ai and elevenlabs.com.

How the programmatic-SEO trick actually scales

Every supported language paired with every other language became its own landing page — drag the slider to see why 60 languages turns into thousands of pages, then pick one to see exactly which pages it generates.

Drag the slider to change language count.
Supported languages60
60 × 59 = 3,540
languages × other languages = landing pages

English alone generates 59 landing pages — one per target language:

SpanishFrenchGermanItalianPortugueseDutchPolishSwedishNorwegianDanishFinnishRussianUkrainianCzechSlovakHungarianRomanianBulgarianGreekTurkishArabicHebrewPersianHindiBengaliUrduPunjabiTamilTeluguMarathiGujaratiThaiVietnameseIndonesianMalayFilipinoChinese (Mandarin)CantoneseJapaneseKoreanMongolianKazakhUzbekAzerbaijaniGeorgianArmenianSerbianCroatianSlovenianLithuanianLatvianEstonianIcelandicSwahiliAmharicZuluAfrikaansCatalanBasque

Keep exploring