DreamOracle: how we took a consumer product from zero to launch
DreamOracle is a dream interpretation platform with an integrated personal journal, built as a progressive web app installable on iOS and Android without any…
By Laurent Maréchal · · 6 min read
There is a fundamental difference between delivering an internal tool to a company and launching a consumer product. In the first case, users are captive — they have no choice. In the second, they can leave in three seconds. That is exactly what this project forced us to internalise at every stage. DreamOracle is a dream interpretation platform with a personal journal, installable directly on a phone, developed for Distr'Action, a studio specialising in audio production and content creation. Here is how it was built, what got stuck, and what you can take away if you are carrying a similar project.
The starting point: a clear idea, an execution to invent
The brief was precise in its intent: create a consumer platform around dream interpretation, with an integrated personal journal and a smooth mobile experience. Not a prototype, not a shaky MVP. A finished, usable, distributable product. This kind of brief looks like classic development work. It is not. From the very first discussions, it became clear that the code was only part of the problem. We also had to decide how the product would present itself, how it would behave on a phone screen without going through an app store, and how it would find its first users. Three topics that most projects push back to later — and that end up killing the launch.
What we built on the technical side
The chosen stack relies on Next.js on the front end and Prisma on PostgreSQL on the data side. This choice is not trivial: Next.js handles both server-side rendering and static pages, which matters for perceived performance on mobile. Prisma simplifies database schema management and migrations, avoiding silent errors when the model evolves.
The most structurally significant technical feature of the project is the progressive web app, installable on iOS and Android without going through the App Store or the Play Store. In practice, the user visits the site, their browser offers to install the application on their home screen, and they end up with an icon and a full-screen experience. It is not a native application, but for the vast majority of use cases the difference is imperceptible. And the distribution gain is considerable: no store validation, no delay, no commission.
We also developed an integrated visual generator that produces images linked to the interpreted dreams. It is exactly this kind of detail that transforms a functional tool into a product people want to share.
Visual identity: what is always underestimated
A consumer product without a coherent visual identity is a product that fails to convince before it has even been tried. We delivered a complete suite in SVG: logo, icons, colour palette, typography, interface elements. Everything in vector format, so that every size is covered without any loss of quality, from the favicon to the loading screen of the installed application.
This work takes time. It requires back-and-forth. And it is often sacrificed in projects that run short on budget or schedule. The result, in those cases, is a product that works but does not reassure the user. In a consumer market, trust is also built visually — especially at first contact.
Acquisition: the part that always gets pushed to tomorrow
This is the topic that almost every project underestimates. The assumption is that if the product is good, people will come. That is not how it works. We integrated thinking about acquisition from the design phase, not at the end of the chain. This influenced technical choices — URL structure, meta tags, loading speed — and design choices — onboarding screens, natural sharing points within the user journey.
The acquisition campaign was built in parallel with development, not after it. This is an organisational constraint, but it prevents you from ending up with a ready product and no plan to make it known. If you are carrying a similar project and want to understand how to structure this chain, our case studies page gives an overview of the projects we have led from end to end.
What got stuck and what it taught us
The main friction point was managing the constraints specific to PWAs on iOS. Apple imposes limitations on certain features of progressive web apps — push notifications, access to certain sensors, cache behaviour — that do not exist on Android. We had to adapt several design decisions to guarantee a consistent experience across both systems, without promising features that would have been degraded on a portion of devices.
The other lesson concerns the visual generator. Integrating image generation into a user flow requires managing variable response times and visible error states. We worked on loading states and error messages with as much care as we gave to the main screens. It is not spectacular, but it is what makes the difference between a product that frustrates and one that people come back to.
If your project involves automated processing or complex flows, our dedicated section on automation details how we approach these topics.
What you can transpose into your own project
DreamOracle can be viewed directly: see the DreamOracle case study. But beyond the product itself, here is what this project concretely illustrates.
If you are building a product aimed at the general public — not an internal tool, not a prototype — you need a complete chain: conception, design, development, go-to-market. These four steps are interdependent. Cutting any one of them weakens the other three. Conception without design produces a product that works but does not convince. Development without thinking about acquisition produces a product nobody finds. Design without solid conception produces a beautiful empty shell.
What teams most systematically underestimate is design and go-to-market. Not the code. Everyone knows code takes time. The other two, people think they can rush through. This project confirmed for us that those are where the essentials are decided.
If your product involves personal data processing or AI features, note that compliance with the European AI regulation is becoming a concrete issue for SMEs: Audityo helps companies with EU AI Act compliance, which is worth anticipating from the design phase onwards.
Conclusion
DreamOracle is live, installable on mobile, and open to the public. This result is the fruit of a complete chain carried out without shortcuts. If you have a product project with the same level of ambition, the best way to understand what that concretely means for you is to talk about it directly. We offer a thirty-minute discovery call, free of charge and with no commitment, to assess together what your project truly requires and how to structure it.