Skip to content
MakerGain
← Back to cases

The screen that feeds you interview answers: its founder made controversy a distribution strategy

Edited by MakerGainPublished: Article updated:

ParakeetAI is a real-time AI interview assistant that listens during calls and provides answers users can read aloud. System-design answers include flowcharts in some output formats; users still speak and exercise judgment themselves. Founder Jure Sotosek treats controversy over AI use in interviews as part of his distribution strategy and acquires customers through short videos. The website currently lists interviews, sales calls and client meetings as use cases.

Imagine a system-design interview. Asked to draw an order-system architecture, you get stuck. ParakeetAI's FAQ says answers to these questions include a small flowchart you can reproduce. The diagram is often omitted in bullet-points mode; Script+bullets or Full script shows it. This is what it hands you during a call: not a voice speaking for you, but a script you can read and a diagram you can draw.[1]

The product calls itself a real-time AI interview assistant, with a website saying it listens and supplies answers in interviews, sales calls and client meetings. Its most conspicuous pitch is invisibility. The site lists five categories: screen sharing, the Dock, Activity Monitor, tab switching and the cursor. Another checking module says “Checked 89 days ago 6 of 6” and invites visitors to open an item and inspect its test. These are the website's own checks, not third-party tests.[1]

Founder Jure Sotosek told Starter Story that reactions were mixed and using AI in interviews was significantly controversial. He considered that consistent with his strategy of making a product that spreads. A founder describing ethical controversy as part of growth says more about the business than any feature page.[2]

It takes over the middle of the call

When creating a session, users choose interview mode or general-call mode for meetings and other conversations. Interview mode reads an uploaded résumé; general-call mode does not. Both use added documents and the Instructions field, which allows twenty thousand characters; longer material is recommended as a document. Users can also message it during a call. It hears questions, reads code in screen sharing and discusses approaches while the conversation continues. Users still have to speak, judge and change the code themselves.[1]

The deliverable is a script to read aloud. System-design questions include flowcharts; other questions do not. Where a diagram cannot be rendered, a text code block substitutes. The desktop app must be up to date to display diagrams.[1]

Image: official product demonstration[3], obtained on 2026-11-18, retaining the original presentation for product illustration only.

A three-day prototype, a weekend MVP, sales within three weeks

Jure says proof of concept took three days: transcribe the call, feed the transcript into ChatGPT and process payment. The stack was NextJS, Stripe, Speechmatics and ChatGPT. Today's product allows selection among Gemini, GPT, Claude and other models without locking models behind subscription tiers.[2] ParakeetAI[1]

The MVP took a weekend and sales began within three weeks. Published January 4, 2025, the interview quotes Jure saying the product launched roughly three months earlier and had just achieved its first about $1,098.90 (€1,000) revenue day.[2]

His spending choices point in the same direction. He did not build the landing page himself, paying an agency about $1,098.90 (€1,000) to design and develop it in Webflow. There was no big launch; acquisition mainly came from short videos on social accounts, with a first sale in the opening week. He says he deliberately chose subjects that could spread naturally—interesting, surprising things people would discuss. Short videos need a reason to stop and talk, and AI in interviews is already controversial. Jure says that fits the distribution strategy, and the interview notes that customers began arriving after videos went viral.[2]

That background does not explain every feature, but it explains a product shaped into a screen suited to short videos. His previous venture, WrumerSound, sold small car accessories online and earned nearly two hundred million short-video views. Having run ecommerce and experienced difficult logistics also informed his move to SaaS.[2] Indie Hackers interview[4]

Weekly, monthly and annual subscriptions—with separate refund conditions

The website's price cards show $78.00 weekly and $149.90 monthly. Yearly is labelled $49.99 per month, “Best value,” with $599.90 billed per year. It says subscriptions include unlimited calls, while credits can also be purchased separately. These are public quotes; actual charges are determined at checkout.[1]

One refund detail deserves attention. A credit-billed session does not count as use if its only credit charge is refunded because you end it within 5 minutes of that charge. Nor does a plan-covered session lasting under 5 minutes, however it ends, when nothing is ultimately charged. Unused purchases can receive a full refund within 14 days; unused subscription renewals within 28 days. For used purchases, a refund may be requested only for ParakeetAI's own technical failure and within 30 days of purchase. Remaining credits do not receive partial refunds.[5] Terms[6]

The product covers interviews, sales calls and client meetings. No public data establishes that all buyers are job candidates. It sells answers to its users; the split between job seekers and meeting participants is not visible in the accounts.

The complete Starter Story title is “My Weekend AI Project Hit $16K/Month In Just 3 Months.” In the interview itself, Jure's milestone is a first about $1,098.90 (€1,000) revenue day. The headline's 16K/Month and the body's about $1,098.90 (€1,000) day are different measures—not MRR, profit or audited current figures—and cannot be substituted for one another or combined into a trend.[2]

A later interview came out on September 23, 2026. Jure said ParakeetAI had gone from zero to more than 1 million dollars in monthly revenue in under two years, with over 1.5 million users. These figures refer to monthly revenue and total users. They give no separate measure of MRR, profit or paying users. [4]

The gap between invisibility and disclosure

The FAQ says the product is intended for conversations where AI assistance is permitted. Any employer may email compliance@parakeet-ai.com to request that its domain or recruitment process be blocklisted. The terms make users responsible for lawful and ethical use, including notice, disclosure or consent where required. Illegal, deceptive and fraudulent uses are prohibited, and access is restricted to people aged 18 or older. ParakeetAI's careers page says candidates can use any AI tool, including ParakeetAI itself, in the company's own interviews. That permission applies to its own hiring process. [1] Terms of use[6] Careers page[7]

There is an observable gap: the product sells invisibility through six checks while assigning responsibility for disclosure to the user. Invisibility can be a product capability while compliance remains a user obligation. Not appearing on screen does not mean the other party has consented. That is a concrete place for criticism: the payer remains responsible for any required notice, disclosure or consent, and the blocklist is a remedy requiring employers to apply rather than a default exclusion. Public materials do not quantify undisclosed use or establish whether it drives purchases.

Among similar tools, HuddleMate says it charges by active assist minutes. Standard is $9.99 per hour, and purchased minutes do not expire. It too listens to meetings and reads the screen to generate answers. Real-time listening and viewing are shared conditions, not mutually exclusive categories. Pricing differs: ParakeetAI sells a subscription including unlimited calls; HuddleMate sells calls metered by active assistance minutes.[8]

Invisibility is a product capability; disclosure is the user's obligation. ParakeetAI sells not being seen while leaving notice, disclosure and consent to the user. Who bears that boundary matters more than how many invisibility tests it passes.

Continue reading

AI productivity
Over 180,000 views, 1 February sale: AI Renamer leaves batch-renaming confirmation to people

Ozgur Ozer's AI Renamer gives content-based names to opaque filenames such as DSC_0412.JPG and scan_20250312.pdf. For manual batches, people inspect the suggested names before pressing Rename All; nothing changes on disk beforehand. Ozer had always given away his small apps and first used Polar to solve payment collection because Turkey had no Stripe support. His first post exceeded 180,000 views but produced 1 sale that month; sales began coming after local support and a Product Hunt launch.

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
It writes your emails, but stops at LinkedIn

According to its website’s workflow, DeepReachAI finds people, researches recent developments, writes messages, classifies replies, and books meetings for cold outreach, while excluding automatic LinkedIn sending, WhatsApp scraping, and calls without consent. In an IndieHackers post, founder Aryan Agrawal said he had expected personalization to be the hook, but the compliance boundary—no automatic LinkedIn sending and no WhatsApp scraping—was what made people connect accounts. The homepage says Orio can send a first email through a connected account, without identifying the account provider.

AI productivity
An assistant cost thousands of US dollars a month. He put AI to work and reached $10k+ MRR

Because an assistant would cost thousands of US dollars a month, Elie Steinbock built an AI email assistant instead. It connects to an inbox, sorts messages, and drafts replies using email history, calendars, and customer records; a person still presses Send. On Product Hunt launch day, he emailed 20,000 newsletter subscribers and posted a video on his YouTube channel with 2,000 subscribers. He estimated that these extra efforts may have added fewer than 50 votes—a judgment, not a measured result.