Oneiric: An Open AI Sci-Fi Film

A 20-minute science-fiction short called Oneiric has appeared on YouTube with an unusually direct label: “100% AI.” The video is presented as an open-source film, suggesting that viewers may eventually get access to more than the finished export. That could mean prompts, storyboards, production notes, editing assets, or a broader record of how the film was assembled. For now, however, the public release is easier to verify than the scope of the open-source promise.

The project is credited to Cinema Studio 4 and was created for the Higgsfield Global Film Festival, a competition associated with a $1 million prize pool. It carries the Higgsfield Originals name, which makes it look partly like a festival entry and partly like a demonstration of what an AI-focused production pipeline can deliver. Those are meaningful signals, but they do not independently prove that every frame was generated without human intervention. The available information should be read as a project announcement, not a complete production audit.

Why a 20-minute film changes the conversation

AI video showcases often focus on short clips: a striking camera move, a convincing character close-up, or a few seconds of atmosphere. Those examples are useful for showing visual potential, but they avoid many of the problems that appear when a story has to keep moving. A 20-minute film has to maintain some level of continuity across scenes, preserve the identity of important characters, support an understandable plot, and survive the less glamorous work of editing. Long-form consistency is where a promising generation model meets the practical limits of filmmaking.

That does not mean Oneiric has solved those problems. A longer runtime can still contain disconnected sequences, repeated visual mistakes, or heavy human cleanup. It does mean the project is testing a more demanding format than a social-media clip. Viewers interested in AI filmmaking should pay attention not only to the most impressive shots, but also to how the film handles transitions, exposition, pacing, and moments where the story requires a character to remain recognizable from one scene to the next.

What “open source” may mean here

The phrase is potentially more important than the film itself. In software, open source normally points to inspectable and reusable source materials. In an AI film, the equivalent is less obvious. A useful release could include the script, shot list, prompts, reference images, generation history, editing timeline, sound workflow, and notes explaining which steps required manual work. A much thinner release might provide only selected prompts or promotional material. Without accompanying documentation, the label remains a promise rather than a practical resource.

• Production assets could help creators study how a story was divided into shots and sequences.
• Prompt and reference records could reveal how the team approached character and visual continuity.
• Editing documentation would show where human judgment shaped the final result.
• Licensing details will determine whether others can legally adapt or reuse the released material.

That distinction matters to independent creators. Someone making a short promotional film, a game trailer, or a proof-of-concept scene does not necessarily need the same tools as a festival production. They do need to know which parts of the workflow are repeatable. If Oneiric eventually publishes a well-organized project package, it could function as a case study rather than just another impressive video. If it publishes only a finished file and a few marketing claims, its educational value will be much lower.

The practical value for AI filmmakers

For creators, the most useful question is not whether AI-generated work qualifies as “real” cinema. That debate quickly becomes abstract. A better question is where the workflow holds up and where human intervention remains unavoidable. Oneiric may give filmmakers a reference point for breaking a longer narrative into manageable shots, planning around inconsistent outputs, and deciding when to regenerate footage rather than repair it in post.

Consider a small team producing a science-fiction pitch for a client or a game concept. They might use a project like this as a planning reference, but they should not assume that a downloadable prompt automatically reproduces the same result. Generation systems can change, source images can affect the output, and many production decisions are difficult to capture in a single instruction. The value is usually in the process: naming conventions, shot planning, continuity checks, and a clear record of what was discarded.

There is also a less comfortable lesson. Calling a film “100% AI” can hide the boundary between generation and production. Story development, selection, editing, compositing, sound design, color work, and quality control may all involve people even when the images or video clips come from AI systems. Until the makers publish a detailed breakdown, audiences should treat the phrase as the project’s stated positioning, not as independently verified evidence about every stage.

What to watch next

The immediate thing to watch is the promised open-source package: where it appears, what files it contains, and what license applies. It is also worth comparing the finished film with any behind-the-scenes material rather than judging the project from its most polished shots. If the release includes enough information to reproduce selected scenes or understand failed attempts, Oneiric will have genuine value for the broader creator community.

For now, viewers can watch the film on YouTube to assess its storytelling and visual coherence. Developers and filmmakers should wait for the underlying materials before treating it as a reusable workflow. Oneiric is an intriguing test of whether AI filmmaking can move from isolated demonstrations toward transparent, inspectable production—and that test has not been completed until the supporting files are actually available.