DomoAI's InkoAI Agent runs 15 AI models in one chat
The Singapore company says the chat assistant writes the prompt, sets format and length, and waits for a creator's approval before every paid generation.

Key takeaways
- DomoAI launched InkoAI Agent on November 13, 2026, a chat assistant that turns one sentence into an image or a finished video clip.
- The agent works across 15 image and video models, including GPT Image 2.5, Nano Banana Pro, Seedance 2.5 and MiniMax H3, DomoAI says.
- InkoAI Agent shows the model and an estimated credit cost and waits for approval before each paid generation, per DomoAI's own launch post.
- DomoAI's release says more than 5 million creators use the platform; the company's About page, read on November 13, says over 6 million users.
DomoAI launched InkoAI Agent on November 13, 2026, a chat assistant that turns one sentence into an image or a finished clip without the creator leaving the conversation.
The Singapore company announced the agent in a press release and opened it the same day at domoai.app/agent. InkoAI Agent writes the prompt and sets the format and length for TikTok, YouTube Shorts or Instagram Reels, and it runs one-click Skills named Short Drama, Anime Video, Cartoon Style and General Story that take an idea to a finished video with dialogue or narration. The creator approves the script, characters and shots along the way.
The agent shows what it plans to generate and waits for approval before it runs, DomoAI says. The company's own launch post adds the mechanics: chat and planning use no credits, and in the default Manual mode every paid generation shows the model and an estimated credit cost before the creator confirms it.
InkoAI Agent works across 15 image and video models, DomoAI says. A Japanese release from the same company lists the split as six image models and nine video models, with GPT Image 2.5 Sunburst as the default for images and Seedance 2.5 as the default for video. Seedance 2.5 is ByteDance's, a 30-second model that accepts up to 30 images, 10 videos and 10 audio files as references. MiniMax H3 is MiniMax's open-weight model, up to 15 seconds at 2K with native stereo audio. DomoAI's release also names Nano Banana Pro and its own DomoAI 2.5.
DomoAI is a Singapore-based image and video platform founded in 2023. Its release says more than 5 million creators use it. The company's own About page, read on November 13, says over 6 million users and more than 40 styles. The count has climbed before: a DomoAI release on May 5, 2026 put it at more than 4 million creators.
InkoAI Agent is not the first video tool to put an agent in front of the work. DeepBrain AI's AI Studios 5.0, released on November 9, 2026, added a Video Agent that writes a script and builds scenes from one line. DomoAI's pitch is breadth: 15 models behind a single conversation, with a confirmation step before each paid generation.
For microdrama, anime and short-form creators, the promise is fewer tool switches and no prompt-writing. Whether the agent keeps that promise depends on the credit math, which DomoAI has not published beyond the in-chat estimate. The independent write-up published by Briefpixel notes that it did not generate a test video and that the launch notice does not establish which actions are free on every plan.
InkoAI Agent is live now at domoai.app/agent; chatting and planning are free, and generations bill DomoAI credits.
Sources
- prnewswire.com - the launch announcement: the capabilities, the 15 models, the four Skills, the 5 million creators figure
- domoai.app - DomoAI's own post: Manual mode, free chat and planning, the credit cost card, the four Skills in detail
- domoai.app - the product page and sign-in entry for InkoAI Agent
- prtimes.jp - DomoAI's Japanese release: the six image and nine video split, the two default models, 7 languages, file limits
- domoai.app - company facts: over 6 million users, 40+ styles, the Elymi companion app
- prnewswire.com - the May 5, 2026 release putting DomoAI at more than 4 million creators
- seed.bytedance.com - Seedance 2.5 is ByteDance's; 30-second generation and the reference inputs
- minimax.io - MiniMax H3: up to 15 seconds at 2K, native stereo audio, open weights
- aistudios.com - DeepBrain's AI Studios 5.0 Video Agent, for the comparison
- briefpixel.com - an independent write-up that says it did not generate a test video
- ai.info - a second independent write-up restating the announced features