Preview

OMEA: Narrative Freedom with Plenty of Potential

After six runs through its only available story, OMEA proves that an AI-powered game master can create a playful, replayable adventure. Punishing dice rolls, long waits and continuity errors show how much the demo still needs to refine.

Introduction

OMEA aims to bring a familiar tabletop fantasy to the screen: describe an action in your own words and watch the world respond. Instead of choosing between a handful of fixed dialogue options, you type what you want to do, the system interprets it and the adventure continues with new scenes, characters and dice rolls.

For this preview, we completed six runs through the only story available in the demo, deliberately choosing different approaches each time. That was enough to confirm the idea’s ability to surprise and encourage repeat play. It also exposed a design that remains too rigid when the dice, narrative continuity or response times fail to support the player’s intentions.

Write the action, accept the response

The loop is deliberately measured. You read or listen to a situation, write how you want to respond and wait for the game to construct the next scene. Some actions trigger a check; others move the story forward immediately. This simplicity makes role-playing feel more natural than searching a dialogue wheel for the closest available answer.

At its best, OMEA makes the same adventure feel meaningfully different when your intentions change. Across six runs, we found enough variation to remain curious about alternative routes. That replayability is currently the demo’s strongest quality, particularly for players who enjoy testing less obvious responses and seeing how a story adapts.

Freedom, however, does not always become genuine control. At several points the story appeared to expect a particular direction, and moving away from it did not produce an equally coherent alternative. The text box lets you suggest almost anything, but the adventure cannot yet turn every reasonable idea into a satisfying outcome.

When the dice take over

The most persistent problem is the severity of its checks. During our sessions, roughly nine out of ten attempts appeared to fail, including actions that seemed straightforward. Failure is an essential part of role-playing and can create memorable consequences, but failure at this frequency discourages initiative: a creative idea stops feeling clever and becomes another opportunity for the dice to say no.

Checks use pools of six-sided dice based on the relevant skill. In our runs, target values were usually five or six, while every roll of one cancelled a success. The system therefore felt stacked against the player even before the story introduced genuinely difficult challenges.

The frustration is amplified by the predetermined demo character. You cannot adjust attributes, build towards a speciality or compensate for a weakness. That creates an awkward contradiction: OMEA invites you to imagine any action, but the tools used to resolve that action have already been chosen for you.

A more permissive balance would not remove risk. Everyday actions could use gentler targets, major failures could be reserved for meaningful challenges, and the game could offer clearer ways to improve the odds. The dice would then enrich the story instead of repeatedly interrupting the fantasy.

A memory that still contradicts itself

Generative storytelling can produce unusual moments, but it can also lose track of its own continuity. In one run we tried to spend the night in a cave. Each time the character slept, the story placed us in the forest; when we returned, the people inside insisted we had never left. One character celebrated our return, then explained away the contradiction as something we had imagined.

The scene was unintentionally funny, but it highlighted a serious weakness for an adventure built around the world remembering where you are and what just happened. The game alternated between incompatible states without reconciling them. Once the player has to ignore a contradiction to continue, the sense of inhabiting a coherent world quickly fades.

These problems matter more here than they would in a traditional branching narrative. OMEA’s credibility depends on preserving locations, relationships and consequences while responding to free-form input. The demo offers glimpses of that continuity, but it needs firmer rules to stop improvisation undoing events established only minutes earlier.

The wait also shapes the pace

OMEA is meant to be a slow reading experience, but a deliberate pause is not the same as waiting for a response to generate. Some scenes take long enough to appear that the momentum disappears, particularly when the previous moment ends with tension or an idea you are eager to test.

Not every response was slow, and the delays did not ruin the six runs, but they occurred often enough to become part of the experience. Faster generation – or a presentation that keeps the scene alive while the next response is prepared – would make the adventure feel far more continuous.

Presentation and language

Text, illustrations, music and voice work make the experience more inviting than a plain chat window. The format retains the calm of an interactive novel while giving characters and key scenes enough presentation to feel staged rather than simply generated.

The demo was available only in English. That is a meaningful barrier in a game where understanding nuance and expressing a precise intention are both essential. The final version will need to show whether it can support additional languages without weakening the system’s interpretation of free-form input.

A promising foundation that still needs work

Even in this early state, OMEA leaves a clear impression: the idea works well enough to make us want to see it grow. A single story encouraged six complete runs, which says a great deal about its capacity to hold attention. It also demonstrates that variety alone is not enough when too many initiatives die on a roll or the narrative loses track of its own events.

The potential is obvious for fans of tabletop role-playing, interactive fiction and patient, text-led adventures. To realise it, AImmersive needs to offer more control over the character, rebalance checks, shorten the waits and strengthen continuity. We are curious to see how far OMEA can develop, but this demo still asks players to tolerate the rough edges of a system in progress.

The good

Different choices produce enough variation to make repeat runs worthwhile
Natural-language input supports improvisation and the feel of a tabletop role-playing session
A promising idea for players who enjoy slower, text-led adventures

The bad

Punishing checks turn too many reasonable ideas into failures
Long generation times repeatedly interrupt the pace
The demo still struggles to maintain narrative continuity

Conclusion

OMEA has the kind of premise that can sustain hours of curiosity: write what you would attempt in a tabletop role-playing game and receive a different adventure in return. Even with only one story available, the demo encourages repeat runs and shows considerable potential. At this stage, however, its dice rolls punish initiative too often, while limited control over character creation and occasional continuity errors restrict the freedom promised by its text box. If AImmersive gives players more agency, balances its checks and strengthens the world’s memory, OMEA could become a compelling take on interactive fiction.

Adventure, narrative RPG and interactive fiction

Recommended for

RPG and interactive-fiction fans who enjoy reading and testing different decisions
Patient players who are comfortable playing in English and curious about generative storytelling