DomoAI announced InkoAI Agent in Singapore on 13 November 2026, bringing image and video production into a conversational workspace. The company says the service can develop an idea into scripts, reference images, shots and an assembled film. For university communications teams, the relevant change is where production decisions and human review occur within that workflow.
What DomoAI announced
The company-issued announcement distributed by PR Newswire describes a creative assistant that writes generation prompts, sets format and length, and carries earlier results forward within a conversation. DomoAI identifies short-form video makers, anime and microdrama creators, original-character designers and marketers as its intended audiences.
The announcement date and the reported service start are distinct. DomoAI’s Japanese release, published on 13 November, says provision began on 12 November. It describes access through a DomoAI account in a desktop browser, with mobile-browser access also available. The international release says the agent is available, placing this development beyond a waitlist announcement or a promised future launch.
The Japanese notice lists six image models and nine video models at launch. Its examples include GPT Image 2.5, Nano Banana Pro, Seedance 2.5, MiniMax H3 and DomoAI 2.5. Those are DomoAI’s descriptions of the models offered through its service; the launch announcement does not establish comparative quality or research suitability.
From a script to separately reviewable shots
DomoAI’s official launch and usage guide describes two paths: individual image or clip requests, and longer productions organised through preset Skills. The four launch workflows cover live-action-style short drama, anime, narration-led stories and cartoon animation. They organise requirements, script development, characters and locations, keyframes, video generation and final assembly.
The guide says creators can approve references before they guide later shots, correct a keyframe before animating it, and regenerate one clip while retaining the other completed work. Earlier takes remain available. Final Cut combines the approved clips for preview and download. The significant workflow detail is that an individual shot remains a unit of revision after generation.
The Japanese announcement also says previous outputs can be redrawn, varied or animated without being uploaded again. A frame taken from a generated or library video can become a reference or the opening image for another clip. These features connect successive production stages around retained material, rather than requiring every request to begin with a new file.
Manual approval and auto-confirm are different modes
The international release emphasises creator approval. The more detailed official usage guide qualifies how that operates: in default Manual mode, a card identifies source material, the selected model, available settings and estimated credits before paid generation. Selecting Generate approves that operation; cancelling returns to the conversation. A paid retry receives a fresh estimate and confirmation.
The same guide says users can enable auto-confirm for the current session, subject to per-card and session limits. Consequently, an account of the launch should distinguish the default approval behaviour from the optional automation setting. A team choosing Manual mode retains a confirmation point for each paid run; enabling auto-confirm changes that arrangement.
For a university team considering the workflow, this is an operational distinction: approval to start a generation and approval to release a university-branded video are separate decisions. The documented generation controls provide places to inspect material, but the institution would still need to decide who signs off on the final communication.
What the early coverage says
AI Ping’s 14 November report highlights the editable storyboard, individual-shot retries and the default Manual mode. It also distinguishes the 13 November announcement from the 12 November service start. Its account draws on DomoAI’s releases and production guide, rather than presenting those features as independently measured performance.
ShorStory Current’s 14 November editorial briefing places InkoAI among developments in short-drama production. It identifies the staged approval structure as the aspect worth examining and explicitly treats the product description as a vendor claim, without claiming to have used the tool. These reports provide editorial coverage of the launch, while leaving output quality a separate question.
Why the workflow matters for university communications
The practical institutional relevance is communications production: turning an already approved explanation into a storyboard, inspecting references and shots, and deciding what may be published. That is an editorial application of the documented workflow, rather than a university deployment announced by DomoAI. A generated scene should be identified as a synthetic image where that distinction matters to its audience.
For researchers and research offices, the useful review question is whether a scene illustrates an explanation or purports to record an observation. A polished generated clip does not by itself establish that a depicted experiment, person or event occurred. The announcement concerns creative production; it supplies no scientific validation for the resulting imagery.
Retaining the source script, approved references, chosen outputs and final editorial decision would make an institutional production easier to explain later. This is a proposed recordkeeping approach based on the agent’s documented reuse of earlier material. It connects to AI provenance, but retained chat results should not be assumed to provide every record an institution needs.
Likewise, a conversation history is distinct from a portable provenance credential attached to a media file. Teams examining C2PA Content Credentials should assess that separately; this launch is not evidence that InkoAI implements that standard. The value for a research office is a clearer place to assign review responsibilities, alongside its existing decisions about disclosure and approved source material.
What comes next
DomoAI says it plans additional Skills for advertisements starting from a product link or photograph, tutorials and AI music videos. The international announcement gives no release date for those additions. They remain planned extensions, separate from the four workflows described at launch. For institutional readers, the immediate development is the available conversational production interface and its documented choice between Manual confirmation and session auto-confirm.
Sources
- DomoAI via PR Newswire, DomoAI Launches InkoAI Agent, 13 November 2026. Checked 2026-11-16.
- DomoAI via PR TIMES, Japanese service launch announcement, 13 November 2026; reports service starting 12 November. Checked 2026-11-16.
- DomoAI, InkoAI Agent Is Live: Make Images and Short Films by Chatting, 13 November 2026. Checked 2026-11-16.
- AI Ping Editorial Team, DomoAI’s InkoAI Agent brings script, storyboard and video into one chat, 14 November 2026. Checked 2026-11-16.
- ShorStory Current, AI short drama brief, 14 November: Methods and workflow, 14 November 2026. Checked 2026-11-16.







