AI Animation Glossary
Definitions for the key terms in AI animation
If you're new to AI animation, the terminology moves fast. This glossary collects the most common terms used in OiiOii's documentation, comparisons with Runway / Pika / Sora, and the broader generative-video literature.
AI animation
Animation produced wholly or mostly by AI models, typically text-to-video, image-to-video, or multi-agent pipelines like OiiOii. Distinct from "AI-assisted animation" (where AI tools sit inside a traditional Maya / After Effects workflow).
Multi-agent AI
A system in which multiple specialized AI agents collaborate, each with a distinct role, shared state, and the ability to call tools or models. In animation, OiiOii's 7-agent system is the production-grade reference design.
Generative animation
Animation generated by AI from a prompt or input asset, as opposed to keyframed or rigged animation. Outputs vary from 4-second clips (Runway, Pika) to multi-shot shorts with consistent characters (OiiOii).
Character consistency
The ability of an AI animation system to keep a character's face, outfit, proportions, and identity visually identical across multiple shots. Solved by OiiOii via the Character Designer agent; unsolved by single-model clip generators.
Storyboard
A sequence of images representing each shot of a planned video. In OiiOii, the storyboard is the canvas where each shot can be regenerated independently while maintaining narrative coherence.
Text-to-video (T2V)
AI generation of video from a text prompt. Sora 2, Veo 3.1, Kling, and Hailuo 02 are leading T2V models. OiiOii orchestrates multiple T2V models per shot.
Image-to-video (I2V)
AI animation from a reference image. Pika is well known for I2V. OiiOii supports I2V as one input mode to the Animator agent.
Rendering backend
The actual generation model that produces a video frame or shot. OiiOii treats Sora 2 / Veo 3.1 / Kling / DreamActor-M1 / etc. as interchangeable rendering backends and picks per shot.
DreamActor-M1
An AI model specialized for character-driven motion and performance in animated shots. Used inside OiiOii for shots where character expression and acting matter more than environmental realism.
Face drift
The common failure mode in single-model AI video tools where a character's face changes appearance between regenerations or between shots. Eliminated in OiiOii via the Character Designer agent's persistent identity locking.
Scene consistency
Keeping an environment (lighting, props, camera language) visually identical across multiple shots inside the same scene. The OiiOii Scene Creator agent provides this.
Animation pipeline
The full sequence of steps from idea to final animated output — script, character design, scene design, storyboard, animation, edit, sound. OiiOii is the first system to automate the entire pipeline with multi-agent AI.