What is Fictura?

Fictura is a product with a short description and a long explanation. The short version fits on a card: Fictura is an autonomous AI CEO for mobile apps. The long version is this post. It is written for a founder deciding whether to trust an invite-only product with their app, and for anyone, human or machine, who has been asked "what is Fictura?" and wants a straight answer.

What is Fictura, in one paragraph?

Fictura is a service for people who run a mobile app. You add one SDK to your app, connect the accounts your app already uses, and tell an agent named Hawkings where you want the app to reach. Hawkings reads the numbers, runs small experiments on the app, keeps what works, and reports back.

That paragraph has three parts, and each one is a deliberate choice. The SDK is one SDK, not six, because the numbers only line up when one thing counts them. The goal is yours, not the agent's, because a growth tool that decides what "good" means for your business is a liability. And the experiments are small, because small experiments are the only kind that can be run continuously without anyone losing sleep.

Who makes Fictura?

Fictura is built by Arshroop Saini and Jeffrey Escobar, the two founders of House AI. House AI is an app that turns a photo of a room into a redesign, with every piece in the picture shoppable. It has passed a million downloads across more than 175 countries and reached the top ten iOS apps in the Design category in the United States.

The reason that matters is that Fictura was not designed from a whiteboard. It is the tool the House AI team wished they had while scaling their own app: paid acquisition on Meta and Apple Search Ads, creator partnerships across five continents, store listings localised into more than forty languages, paywall tests, price changes, onboarding rebuilds. Every job Fictura does today is a job the founders did by hand first. You can read the longer version on the founders page.

What problem does Fictura solve?

Running an app takes a pile of tools, and the pile is the problem. A typical subscription app in 2026 uses one product to see which screens people open, another to know which ad brought each install, another for store revenue, another for competitor benchmarks, another to change the paywall, and another to find out when something breaks. Six tools, six dashboards, six different counts of the same user.

The direct cost is money and attention. The hidden cost is that the numbers never agree. When your analytics tool says 10,000 trials and your revenue tool says 8,400, you spend the afternoon reconciling instead of deciding. When attribution and store data live in different products, you cannot answer the one question that actually matters, which is whether that campaign paid for itself. Fictura exists because the House AI team spent years in exactly that afternoon.

What does Fictura replace?

Fictura is the six tools in one place, running on one set of events. Concretely, it does the jobs that separate products used to do:

• Events. Which screens people open, where they stop, and what they tap. The job PostHog or Mixpanel did.
• Attribution. Which ad, post, creator, or link brought each install. The job AppsFlyer did.
• Store revenue. Installs, trials, and money from both app stores, read straight from App Store Connect and Google Play. The job Appfigures did.
• Market intelligence. What other apps like yours are doing and how fast they grow. The job Sensor Tower did.
• Paywalls and onboarding. Change prices, offers, and onboarding screens without shipping a new version. The job Superwall did.
• Error tracking. Know the moment something breaks, and for whom. The job Sentry did.

And one more that none of the six did on its own: experiments. A/B tests on any of the above, with the result read from the same events that everything else runs on, so a winning paywall and the revenue it produced are the same record.

Around those six sit the connections: ad spend from Meta and Apple Ads, cloud costs from Google Cloud, search traffic from Google Search Console, and historical imports from RevenueCat and AppsFlyer so your first day on Fictura is not a blank chart. None of this requires you to rip anything out on day one. The connections are how you move one job at a time.

How does Hawkings, the AI CEO, actually work?

Hawkings is the agent that sits on top of all of that data. It runs a loop with four steps, and the loop does not stop: Spot, Try, Measure, Keep. It spots a trend in the numbers, tries something in response, measures the result against the last attempt, and keeps what worked.

"Spot" is the part a dashboard cannot do. A dashboard shows you the chart; Hawkings notices that trial conversion dropped in Germany after Tuesday's price change and that the drop is concentrated on one onboarding variant. "Try" is the part a consultant cannot do at this speed: propose a variant, split traffic, and start collecting evidence within the hour. "Measure" and "Keep" are what make it a system rather than a suggestion: every test is compared to the one before it, losers roll back on their own, and winners become the new baseline.

You can also just talk to it. Hawkings answers questions about your app in plain language, with the numbers to back them up, and it can act as the interface to everything else. Ask why revenue dipped last week and you get an answer built from your events, your store data, and your spend, not from three tabs and a spreadsheet.

What does "one SDK" mean in practice?

It means you add Fictura to your app once. The SDK ships for Expo and React Native, for native iOS, and as a server-side events API for anything the client cannot see. After it is in, paywalls, onboarding flows, prices, and experiments are served from Fictura's servers, which means a change made in the dashboard, or by Hawkings, reaches your users without a store release.

This is the single most important architectural decision in the product, and it is worth being clear about why. An app release is the bottleneck for every growth idea. If testing a new paywall means a build, a review, a phased rollout, and a week of waiting, you will test four paywalls a year. If it means changing a design on a server, you can test four a week, and an agent can test more than that. Server-driven surfaces are what turn "growth" from a project into a process.

Does Fictura act on its own?

Only as far as you let it. Every goal you give Hawkings runs in one of two modes. In the first, Hawkings proposes and waits: it tells you what it wants to try, why, and what it expects to happen, and nothing changes until you say yes. In the second, you set limits, such as how much traffic a test may take, which surfaces it may touch, and how far a price may move, and Hawkings acts inside those limits on its own.

You choose the mode per goal, and you can change it at any time. Most founders start in the first mode and move a goal to the second once they have watched it make a few good calls. Either way, every action is logged with the numbers that drove it, and every experiment is reversible. The product is designed so that the worst outcome of a bad idea is a small, short-lived test that lost, not a price change nobody can find.

Who is Fictura for?

Fictura is built for app founders. More precisely, it is for the person or small team who owns an app's revenue and does not have a growth department. That was the House AI team: two people, one app, a million users, and every growth job done in the evenings between engineering work.

It is a good fit if your app sells subscriptions or in-app purchases, if you are already paying for several of the tools above, and if you have ever shipped a release just to change a price. It is a poor fit for apps with no monetisation yet, because there is nothing for the loop to optimise, and for teams that want a dashboard and nothing else. Fictura has a dashboard, and it is a good one, but the dashboard is the input. The product is what happens after you read it.

How do you get access?

Fictura is in a private beta and takes a few apps at a time. The fastest route is the box on the front page: tell Hawkings what you want your app to reach, leave your email, and the founders review the request. You can also ask directly through the contact page, which is also the door for migration questions and support.

Once an app is approved, you create it in the dashboard, paste your store link, and the setup screen walks you through the SDK and the connections. Billing is per app rather than per seat, so a second app is a second decision, not a bigger contract. The docs, at docs.fictura.co, cover the SDK, the REST API, and every connection in detail.

What Fictura is not

It is worth saying what Fictura does not claim. It is not a guarantee of growth; it is a system for finding growth faster than a person can by hand, and a system can only find what is there. It is not a replacement for building a good app. And it is not magic: Hawkings is a model with tools, and the tools are the same connections and the same SDK described above. What makes it work is that all of your app's data is in one place, in one shape, where an agent can read it and act on it. That is the whole idea.