Redefining Detective Games: Inkblood, Moody Deduction-Led Occult Puzzle

Inkblood has been described as a cozy occult detective game, which is a very specific and compelling combination. How did the team arrive at that tone, and what does “cozy” mean inside a story about murder, cults, and supernatural investigation?

Petr Škorňok, HeyBird! co-founder, CEO, and producer: To be honest, the term "cozy" surprised us at first. It wasn't exactly our conscious goal, but when players and streamers started using it to describe the game to each other, it actually clicked for us. From the very beginning, we were drawn to games that manage to build a strangely intimate – almost homely – atmosphere despite a dark setting.

We really loved the idea of creating a tangible place that the player can make their own. You put things down, organize your small interior, and discover its hidden layers. It serves as your cozy, familiar, safe haven while truly terrifying things are happening out there, behind the glass. But as development went on, we realized we wanted to take it a step further. We wanted to build up this sense of perfect safety specifically so we could slowly start to disrupt it later, letting the outside danger quietly seep in.

We also simply didn't want jump scares in this game. We are much more interested in building that thick atmosphere of occult and folk horror that slowly gets under your skin and stays there.

The game is built around solving mysteries without being told exactly how to solve them. How do you design an investigation system that trusts the player’s deduction while still giving them enough structure to avoid frustration?

Petr Škorňok: The principle we keep coming back to is that the player should be uncertain about the answer, never about how the game works. Early on, we made the mistake of designing single points of failure. One clue led to one answer, and if you missed it, you were stuck. Playtests showed us time and time again that what we thought was a clever deduction puzzle was actually just forcing players into desperate pixel hunting. We fixed this by making sure several clues point toward the same answer. It stops the player from scanning the screen for one magic object, and when they have already formed a theory, they get a chance to confirm it instead of just guessing.

When players do hit a wall, we rely on Agatha. She is the Inquisition's most devoted fan, a character who follows your investigation and helps you out through a small window at the back of the carriage, and she is also our hint system; the Watson to your Sherlock. Every great detective needs a sidekick; it's a classic genre trope.

We wanted a personalized hint system that felt alive, rather than a generic question mark button in the corner of the UI. She gives tiered hints, letting you decide how far down the ladder of help you want to go without ever breaking the fiction.

Finding the exact line between guiding the player and fighting them took a long time. There is a specific case in the first third of the game that we iterated on endlessly. It never quite clicked until much later in production. Once we had designed enough of the later cases to finally understand what our "just right" difficulty felt like, we took all that newly acquired experience, went back, and completely rebalanced the first half of the game.

Players inspect clues, take photos, build case files, and call their overseer once they have a strong theory. Can you walk us through the design process behind those investigative tools and how they work together?

Petr Škorňok: We designed the investigation as a sequence: observe, capture, organize, interpret, commit. Every tool in the carriage supports a specific part of that loop.

The camera represents the first real decision you make. You can photograph any pixel in the game, which means a clue doesn't become evidence just because you clicked on it; taking the photo is you saying, "I think this matters." Because this level of freedom isn't very common in the genre, it can actually be a bit confusing for players at first. But once they find their rhythm and learn how the game "thinks," it clicks beautifully. We did have to iterate on things like photo limits so the inventory wouldn't become a mess, but honestly, we are just excited to see the weird and wonderful things players will photograph and piece together.

When it comes to physical objects, those are brought in by the magpie. Our studio is called HeyBird!, so slipping a bird into our games is kind of our signature. In our previous game, Ministry of Broadcast, it was an annoying, talkative crow. The magpie in Inkblood doesn't really get to speak, but it felt comforting to have a living animal companion in the carriage. Plus, magpies are famously known for stealing and hoarding things, so having her fetch physical evidence from the outside world just made perfect sense. Everything she brings in ends up on what we call the altar. It is a surface in the carriage where you can lay out the objects you have collected and physically arrange them next to each other. That sounds simple, but it is where most of the actual thinking happens. Photographs and letters tell you things in isolation, and the altar is where you start putting them into relationships. It is also the part of the interior that changes the most over the course of a case, which is a big part of why the carriage feels like your space rather than a menu.

Then there is the communication barrier. Apart from the coachman, who has a map tattooed on his back, everyone else communicates with you through letters handed through the glass. We wanted to capture the sensation of being a bureaucrat behind a counter, which fits the austere, administrative nature of the Inquisition in our world. But looking back, it also heavily channeled our own experiences from the COVID era – communicating with the outside world through a pane of glass, keeping a physical barrier between you and others, and harboring this lingering anxiety about direct contact. It makes the carriage feel safe, yet slightly inhuman.

The last step is committing to an interpretation, and that is what calling your overseer is for. The game deliberately doesn't confirm every small deduction along the way. We let uncertainty accumulate instead, so at some point you have to decide for yourself that you understand the case well enough to say it out loud. That moment of "I am fairly sure, but not completely sure" is exactly the feeling we wanted to protect. If the game validated every step, there would be nothing left to be brave about.

The supernatural magnifying glass lets players peer into the past, turning time itself into part of the investigation. How did that mechanic evolve, and what kinds of puzzle or storytelling opportunities did it create?

Petr Škorňok: Because Inkblood deliberately doesn't rely on massive walls of text, we had to lean heavily into visual storytelling. We needed a system where the player could directly compare the aftermath of an event in the present with what actually happened before. That's where the Echo Glass – our peephole into the past – made perfect sense. The magnifying glass got interesting the moment we started treating the past simply as another layer of physical evidence.

An object might be gone in the present but visible before. A scene can tell a completely different story depending on when you look at it. You can hold what somebody claims happened against the visual traces of what actually happened. It also allows absence to become evidence. Sometimes the biggest discovery isn't that something was there in the past, but realizing that something you expected to be there simply isn't. We didn't need a new puzzle language. You are still observing and comparing; you just have access to impossible information.

From a production standpoint, making two versions of an environment wasn't actually a coding nightmare. The real challenge was on the asset and design side to keep everything consistent across timelines. Tracking exactly what state an object should be in, and when, was one of the main reasons we had to build an absurdly large digital board to map out the entire game.

Inkblood’s cases are described as interconnected rather than isolated. How do you plan a mystery structure where a small clue from one case can later become important to the larger conspiracy?

Petr Škorňok: Clues operate on two levels. A piece of evidence has to work logically in the specific case where you find it, but some of them, like a name, a symbol, an organization, play a much larger role that might just seem like set dressing at the time. When something similar surfaces later, your entire reading of the original case changes. That delayed recontextualization is what we enjoy building the most. We want you to slowly realize that supposedly unrelated cases share the same fingerprints.

Practically, managing this means relying on a massive Miro board. If I open it today, honestly, the sheer scale of it is terrifying. It contains two and a half years of interconnected threads and details. We keep it highly organized, but the pure density of information makes it incredibly overwhelming for anyone new joining the team. Every case has its own smaller section, and above all of them sits this master board tracking which thread connects to which across the entire game. It's the only way to see if a storyline actually pays off or just quietly dies somewhere in the middle.

The biggest danger we faced was turning these overarching connections into a memory test. If a major revelation depends on you recalling one tiny object from five hours ago, that doesn't make you feel clever; it just makes you frustrated. We absolutely ran into this during playtesting. We would think a connection was obvious, but players had completely forgotten the setup. Because of that, we constantly have to go back and retroactively plant more clues and echoes throughout the earlier parts of the game to make sure those connections actually land. We want the conspiracy to feel like it was there from the start, not like something we just tacked on in the third act.

From an art direction standpoint, how did you define the visual language of Inkblood’s world, from Inqburg and abandoned villages to swamps, rite sites, dead bodies, occult details, and the carriage itself?

Petr Škorňok: Our art director is Natalie Konecna. When we were figuring out the visual language, a shared reference point the team already knew and loved was a 1981 Czechoslovak film called The Mysterious Castle in the Carpathians. It does something we find incredibly difficult to pull off: it holds humor, grotesquerie, folklore, and genuine seriousness in the exact same frame without any of them cancelling the others out. We also drew heavily on the tension of the turn of the 19th and 20th centuries and the massive gap where you could travel a day's distance from a technologically advanced city and arrive somewhere people still firmly believed in ghosts and myths.

Visually, the game has a printed, slightly degraded quality thanks to dithering and textures that lean toward old printmaking and risograph. We wanted something tactile and imperfect instead of a clean digital surface. But honestly, making a detective game look "dirty" is dangerous because readability is everything. We had to fight for that balance. To solve this, we had to selectively increase the level of detail in key areas, and we even introduced a pair of binoculars. This lets players zoom in on specific spots in a scene where we needed to convey precise visual clues that would have otherwise been lost in the stylistic noise.

The carriage acts as the visual counterweight to all of it. Every location outside has to function as a complex investigative space where composition can't just be decorative. You stare out at an abandoned village or a ritual site in the forest, analyzing every pixel, and then you pull back into this cramped, cluttered room full of your own familiar tools. That constant visual contrast is really the core identity of the game.

What does your production pipeline look like for building a 2D investigative adventure? What tools and workflows do you use for environment art, character art, UI, animation, puzzle scripting, writing, and implementation?

Petr Škorňok: We are a team of five full-time people and five freelancers, so our pipeline has to be strictly gameplay-first. A mystery might look beautiful on a design doc, but it can still be absolutely miserable to actually solve.

Everything starts on our massive Miro board. We map out the cross-case connections, then build a smaller board for the specific case to track exactly what the player knows at every stage. From there, we build a graybox in Unity and playtest it. Only when the logic works do we move to final art, which is done in Photoshop using our own custom brushes.

That order matters a lot. Moving from a graybox to final art can completely break how a player reads a scene. It happens to us constantly – a clue that was perfectly clear as a gray block suddenly vanishes into a beautiful, detailed environment, or conversely, a newly rendered object shouts at the player long before it should. Because of that, the case has to be playtested again, several times, after the art lands. Animation comes last and we do it in TVPaint. It is also what dramatically elevates the overall quality of the game. Our lead animator, Ivan Strobl, who also handles parts of our trailers, brings an incredible amount of life to everything we build.

We rely heavily on diegetic UI, where the camera, the evidence, and the deduction board all feel like real objects, which means writing and puzzle design can't be separated either. Dialogue, documents, and physical clues form one very fragile information network. We do have constant, highly detailed discussions about swapping out just a single word. Replacing one specific word in a letter can drastically shift how a player perceives a clue and completely alter the difficulty of a deduction.

The game draws comparisons to deduction-focused titles like Return of the Obra Dinn and The Case of the Golden Idol, but it also has its own gothic, supernatural identity. What lessons did you take from those kinds of games, and where did you intentionally go in a different direction?

Petr Škorňok: We absolutely love games that make the player feel smart, and both of those do it perfectly by giving you information and then refusing to perform the final act of interpretation for you. But we also looked at them from a production standpoint. I remember listening to an interview with Lucas Pope about Obra Dinn, and one major takeaway for us was that we really didn't want to build a deduction system based on stringing together words and grammar, because that quickly becomes a nightmare for localization. Instead, we wanted to take that satisfying feeling of deduction and mix it more with tactile, escape-room-style mechanics.

We also went intentionally in a different direction when it comes to the player's role. In games like those, you are essentially a chronicler – an external observer piecing together events that are already over. In Inkblood, while you aren't making branching narrative choices that change the story, you are the main character actively driving it forward. You are a direct catalyst for what is happening, not just a bystander documenting the past.

And then, of course, there is the carriage. It cost us an enormous amount of effort, because the whole game rests on it, but we never once considered dropping it. It was a day-one design decision by our lead game designer, Dominik Konecny. His original pitch was that the investigator is simply too important to ever step out into the mud. We eventually evolved the narrative reasons for why you stay inside, but that core mechanical limitation stayed. You investigate entirely by observing what happens outside and by deciphering mechanisms, letters, and objects inside, which creates a completely different kind of immersion.

For 80 Level’s audience of artists and developers, what has been the hardest part of making a mystery game where the story is not simply delivered to the player, but assembled through observation, experimentation, and systems?

Petr Škorňok: Honestly, the hardest part is getting inside the minds of the players. We are constantly trying to design an experience that makes them feel clever, which means pacing the clues carefully so they can sense a twist coming, rather than having it feel completely random.

Because of this, playtesting is absolutely everything. We don't just watch to see if someone solves a puzzle. We watch how their mental model develops. We look for what they noticed, what they ignored, the hypotheses they formed, and the exact moment something finally clicked. And if they got it wrong, we have to figure out if it was an interesting misunderstanding or just a failure of missing information.

What constantly surprises us is how differently players are wired. Some of our puzzles lean on exploration, others on pure logic. A puzzle one player flies through might completely block another, and on the very next challenge, their roles will flip. When testers give you completely opposite feedback on the same section, you can't just listen to one of them. You simply have to run more and more playtests until you find that average, middle-ground value. There is no single perfect difficulty curve, only a median of very different experiences.

Another constant tension is side content. We love putting flavor characters and small side stories into the world to make it feel lived-in. Content that many players will never even see. But players on the critical path will scrutinize every character and every single pixel for a solution. So, a tiny detail that wonderfully enriches the world for one player might actively mislead another.

This affects every single discipline on the team. An artist can make a clue slightly too prominent. A writer can give away one word too much. An environment can be so rich that the important thing gets lost in the noise. You ultimately end up designing the "absence" of information just as carefully as the information itself. But when it all finally works, you get something conventional storytelling just can't do. The player doesn't learn what happened because a character told them; they feel like they discovered it themselves. Protecting that sense of ownership over the conclusion is exactly why we do it.

Petr Škorňok, HeyBird! co-founder, CEO, and producer

Interview conducted by David Jagneaux