Skip to content
MakerGain
← Back to cases

Six models in one chat: Magai sells not having to explain it all again

Edited by MakerGainPublished: Article updated:

Magai calls current models from multiple labs on users’ behalf. Its website says users can switch inside one conversation, with memory shared across models so they can pick up where they left off. Founder Dustin W. Stout says he cut five pricing tiers to two and removed the free trial; pricing lists a 30-day, 100% money-back guarantee. While receiving conference-booth quotes, he chose a 90-minute podcast appearance and says the business nearly doubled in two weeks. He left the cause open.

Magai's homepage runs a product demonstration: GPT makes an outline, then /Claude write an email writes an email, /Perplexity Sonar checks competitor pricing, /Seedream makes the main image, /Gemini turns it into a video script, and /Kling animates that image. Six steps, one chat box, without breaking the conversation.

This is an official demonstration, not an independent test. But it makes the problem clear. Ordinarily, that sequence takes several tabs and subscriptions, with the background explained again whenever the tool changes. Another outcome is easy to imagine: rather than switch, someone uses whichever model is at hand for everything.

Magai's role is specific. Its FAQ says it calls the labs' own APIs on users' behalf, providing current models from OpenAI, Anthropic, Google, xAI, and DeepSeek, rather than replicas or cheap substitutes. Type a slash and a model name to switch; the selected model generates the next response with the whole conversation available. Memory sits above the models and follows the user across switches. Users can read, search, and delete it. [1]

Image: official product demonstration[1], obtained on 2026-11-18. An actual public homepage demonstration, not a personal test or evidence of business revenue.

Cutting five prices to two

Founder Dustin W. Stout describes a decision in his own account. Magai originally had five tiers at $9, $19, $39, $79, and $99. He cut them to Solo at $20 a month and Team at $40, while removing the free trial. He says that brought better customers, higher conversion, and easier comparison. [2]

Easier comparison deserves a pause. Five prices make users answer a multiple-choice question before starting; two tiers leave two options. That need not be better for everybody, but at least it makes the initial choice quicker.

Without a free trial, what happens to risk? As of November 18, 2026, pricing promises a 30-day, 100% money-back guarantee without requiring a reason. Trying the product moves from before payment to after it. The exchange becomes “pay first, refund if dissatisfied,” demanding more trust and resting on a refund promise on the pricing page. [3]

As of November 18, 2026, the price list has three tiers—$20, $40, and $200—marked as billed monthly. Standard says “All Standard Models + Tools,” while Pro and Ultra say “All Premium Models + Tools.” The help page, however, says every plan includes all models, including premium ones. The pages still describe the main differences as monthly usage, scheduled-task counts, and support priority. Public pages do not give a consistent answer on Standard's premium-model access. The same screen says “Save 17%” without specifying the relevant billing cycle or an annual monthly equivalent. Additional Standard users cost $20 each; Pro and Ultra use custom pricing for added users, and more than 10 seats go through enterprise or volume sales.

A meter behind “one price”

“Every top AI. One chat. One price.” is the homepage slogan. The help page unpacks it: monthly usage is a credit balance matching the fee, approximately $20 for Standard, $40 for Pro, and $200 for Ultra. Messages, images, and videos reduce it using model multipliers: 0.1× for low-cost models, 1× for standard, and 3× or higher for premium models. After exhausting the allowance, users can top up $5, $10, or $20 anytime, without a cooldown. [4]

In other words, $20 does not buy $20 of top-model conversations. It buys $20 of purchasing power, with different models priced differently. This is not concealed—the help page is explicit—but it shapes the business. A simple monthly fee at the front is backed by multiplier-based metering. That is the aggregator's real calculation, rather than the length of its model list.

These are the website's claims. Switching smoothness, context preservation, and image and video quality have not been independently tested.

Four years, and the bugs in version one

In a blog post dated August 12, 2025, Dustin W. Stout says he spent four years before Magai building products nobody wanted, through pivots, failed launches, and nearly giving up. The post says Magai reached $1 million in cumulative sales at 32 months old and was then approaching $100K MRR. Those are founder-reported figures, not profit or audited revenue. Gross profit, costs, refund rates, and paying-customer counts are undisclosed.

In the same account, he describes a buggy, slow first version with a clumsy interface and inconsistent performance. It nevertheless solved the core problem well enough to attract users. Personas, workspaces, and team collaboration followed repeated customer requests. [2]

This should not become a tale that persistence guarantees a return. He does not explain in detail what went wrong with the products over those four years, nor can we assert a causal link to Magai's later pricing decisions. What is clear is the sequence: Magai did not start with a complete architecture and then search for users. Many features were demanded by users.

He recorded ninety minutes instead of buying a booth

Another choice he repeatedly discusses concerns acquisition.

Around the time he received several $30,000–$40,000 invitations to sponsor conference booths, he chose a 90-minute Brad Lea podcast interview instead of a booth. He had already appeared on more than 100 podcasts. He says that episode nearly doubled the business in two weeks, without providing verifiable incremental revenue, an attribution method, or interview production and distribution costs. The $30,000–$40,000 figures were quotes, not money necessarily saved.

He describes repurposing the interview for several platforms: a full YouTube video and clips for Instagram, Facebook, TikTok, and others. Dustin W. Stout's account[5] Putting booth quotes beside podcast results can easily become a zero-cost growth story. But quotes are not invoices, and doubling lacks attribution.

Two bets in the same market

ChatHub bets on another experience: ask one question and have ChatGPT, Claude, and Gemini respond simultaneously, comparing the answers side by side. Its website claims support for more than 20 models and over 300,000 users, both self-reported. Pricing offers monthly payment or 35% savings with annual payment. Subscriptions include model access without separate API keys, with monthly allowances by plan. [6] · [7]

The products overlap, but their default user actions differ. One wants you to see disagreements; the other wants you to stay in the conversation. Which is worth more depends on whether users are choosing an answer or completing a task. One shortcoming cannot settle the choice for every customer.

A short promise

Follow the thread and Magai sells a short promise: switch models without explaining everything again. The $20 monthly Standard tier supplies its starting purchasing power. Its difference from another subscription involves more than price: context travels with the user, alongside images, video, agents, and integrations bundled into the subscription.

The shorter the promise, the harder the accounting. “One price” works at the front because a 0.1×, 1×, and 3× meter supports it behind the scenes. Magai's craft lies in that memory layer and billing system, rather than the model list. The simpler the front, the more important it is for the metering behind it to work.

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.