Glossary
The words Ordinary Animator uses -- what each one means, and how they fit together.
The vocabulary Ordinary Animator uses. The first section shows how the main words fit together; the list after it explains each one, alphabetically. Where our word differs from the industry's, the entry names both.
How it fits together
Your show is a project. Everything else lives inside it.
- The story nests: a project holds episodes; an episode holds scenes; a scene holds shots. A shot is one clip.
- The cast and sets sit beside the story, not inside it: characters, locations and props belong to the project and are reused in every episode. Each can have looks. A shot points at the characters, location and props it shows, and picks which look of each.
- The screenplay drives the story: an episode's script -- its scene headings, action and dialogue lines -- becomes the scenes, the shots and the audio. Each dialogue line gets one or more takes.
- The plan says what to do: each episode has an Episode Plan of phases, steps and criteria. The guide bar on every page reads it and says what comes next.
How a shot gets made
- A visual style says what a scene means visually ("Normal", "Dream scene"). It is yours. A scene, a shot or a look is put in one. A starter style gives you one ready-made.
- Each phase makes one or more media purposes -- storyboards, rough clips, finished clips, references. Under each media purpose, the style holds an ordered list of recommendations.
- A recommendation is a recipe plus an art style for each model the recipe uses. The first recommendation that fits the shot is used; if none fits, the shot says so.
- A recipe is a series of operations. Each operation is a ComfyUI workflow or a built-in operation. A workflow runs one model, at one build.
- An art style is what one model draws ("Flat cel anime"). It belongs to the model, not to you, and it supplies the prompt wording -- a visual style has none of its own.
- Sound has no look, so it is not under a visual style: it has one list of recommendations for the whole show.
- What comes out lands in the shot's media gallery. You choose its storyboard, then its rough clip; the Finish makes its finished clip. The Sequencer plays the episode from them.
Trying things out
- An experiment tries several recipes on your own artwork. When you Decide, the winner goes to the top of the style's recommendations.
Help along the way
- The AI Assistant answers questions and does work for you, using skills. It suggests; nothing changes until you accept.
A to Z
Agent -- the older name for a configurable specialist behind the Assistant, with its own instructions, context and tools. Skills replaced them. Agents you configured before that are still listed, and still editable, in the project's Skills view; new ones are not created.
AI Assistant -- the AI helper you type to (or speak to, by holding the microphone) to ask questions, navigate, and get things done. You have one Assistant, and it reaches for its own skills as your request calls for them. See The AI Assistant.
AI Models -- the model catalogue, in your account menu: every model there is, what to read about each one, and whether its files are on this machine. Nothing on it belongs to a project. See AI Models.
Arc -- a thread that runs through your series and develops over time: a season's dramatic question, a character's growth, how a location changes, how a relationship shifts. You place an arc on the Timeline by drawing it across the episodes it covers.
Art style -- what one AI model actually draws ("Flat cel anime", "Soft painterly"). It belongs to the model and ships with Ordinary Animator; on each recommendation you choose one for every model the recipe uses, and that choice supplies the prompt wording. A model with no art styles listed uses "the model's default art style". Not the same as a visual style, which is yours and says what a scene means. AI Models shows each model's art styles. See Visual styles.
Build -- which file of one model you run: full size, a lighter fp8, a quantized GGUF. A build is a hardware choice (memory against a little detail), never a look. See Models, builds and LoRAs.
Built-in fallback -- the ready-made recipe that runs for a media purpose when its visual style has no recommendations, so the show still renders.
Built-in operation -- a small, single-purpose step the platform runs itself (text-to-speech, assemble, trim, voice changer, ...), as opposed to a ComfyUI workflow. (Not "virtual workflow".)
Chain -- several operations run in order, each one's output feeding the next. A gallery's Recipe panel shows a chain as the steps of its recipe, and Run all runs them. What you see of the chain afterwards is the arrows in Graph view. See The AI Assistant.
Character -- someone in your show. Characters belong to the project, not to an episode, so the same character appears in any number of episodes. Each has Story properties (role, backstory, personality), Appearance, a voice and looks. See Characters.
ComfyUI -- the AI generation backend. It runs either locally (your machine, via the Chrome extension) or on Comfy Cloud (managed GPUs). See ComfyUI Integration.
Criterion (acceptance criterion) -- one thing that must be true before a step of the plan is done ("every speaking character has a voice"). You sign it off; it is approved or it is not. See The Episode Plan.
Decide -- how an experiment ends: you say which result won, and its recipe goes to the top of the visual style's recommendations for that media purpose. See Experiments.
Dialogue line -- one line a character speaks in the screenplay. Each line becomes audio through one or more takes, and a shot plays the lines that fall inside it. See Dialogue audio.
Draft voice -- the voice a character is set to use by default, marked with a star, that one-press dialogue generation uses for every line.
Entity -- anything Ordinary Animator tracks with a type, an ID, and properties: a project, episode, scene, shot, character, location, prop, and so on. See Entities.
Episode -- one instalment of your show, with its own screenplay, scenes, plan and Sequencer. It uses the project's characters, locations and props. See Episodes.
Episode Plan -- the checklist each episode works through: the phases, broken into steps, each with its criteria. It records what is needed by what, not a fixed order. The guide bar reads it to say what to do next. See The Episode Plan.
Experiment -- a question with an answer: a disposable workspace, with its own Media gallery, where you try models on your own artwork and then Decide which won. It belongs to the project and is about one or more subjects. See Experiments.
Finished clip -- the full-resolution clip the Finish makes from a shot's rough clip (or that you set yourself with Set as finished clip). It plays in place of the rough clip, which stays chosen beneath it. See the Finish.
Fountain -- the plain-text screenplay format the Script is written in. See fountain.io.
Guide bar -- the bar across the top of a page that says what this page is for, lets you sign it off, and says where the plan goes next, with a Go button to take you there. It follows the Episode Plan. See Navigation.
Idea -- a note in the project's scratchpad: a thought to come back to, not yet part of the story. See Ideas.
Identity LoRA -- a small add-on file trained on one character, location or prop so it looks like itself in every shot. It is made for exactly one model, and it is the only kind of LoRA you handle yourself: you upload it into the subject's look gallery. See Characters.
Location -- a place in your show. Like characters, locations belong to the project and are reused across episodes, and can have looks ("at night"). See Locations.
Look -- an alternate version of a character, location or prop ("winter coat", "at night"). It keeps the subject's Story properties and can override its Appearance. A look can have its own visual style. A shot picks which look of each character it uses.
Media gallery -- the shared pin-board of images, clips, and audio for an entity. Starring marks an item as starred, a bookmark that floats it to the top; archiving sets it aside. See The Media Gallery.
Media purpose -- what a phase makes, and what the media is for: character, location and prop references, scene images, storyboards, rough clips, finished clips. Fixed by Ordinary Animator. A visual style holds its recommendations under each one. (A media type is video, image or audio; a purpose is what the media is for.)
Model -- a single AI model that generates media, known by its name (Wan I2V 2.1, Flux Klein 2.0 4B) and published in one or more builds, each with its own hardware cost. A workflow runs one model at one build. You do not pick a model on its own: a recipe names the models it uses, so the model comes with the recipe. Each model has its own art styles. See Choosing the right model.
Operation -- one unit of work with named inputs: either a ComfyUI workflow or a built-in operation. You choose one in a gallery's Recipe panel, fill it in, and run it. See The Media Gallery.
Performance -- your own acting, face or voice, recorded or uploaded to drive a shot. Each shot with characters asks whether it is driven from a performance; the episode's Performances to record page lists every shot still waiting for one.
Phase -- one of the ten stages of the production pipeline: The Foundation, The Cast, The Story, The Script, The Sound, The Visuals, The Boards, The Shoot, The Finish, The Final Cut. Each is our plain handle for a real studio stage. A phase makes one or more media purposes. See The Production Pipeline.
Project -- your show: the top-level container that holds every episode, character, location, prop, experiment, idea and visual style. You own the projects you create and can share them. See Projects.
Prompt guidance -- per-model notes on how a given model likes its prompts: some want a short list of keywords, others a long, detailed description. Ordinary Animator applies a model's prompt guidance for you when it turns your captured properties into a prompt, so the wording suits whichever model the step will actually run. The notes are researched and shipped per model -- there is nothing to set up. See Models, Workflows & Recipes.
Prop -- an object that matters to your story. Props belong to the project, are reused across episodes, and can have looks. See Props.
Property -- a named field on an entity. Core properties are its primary fields (name, description, ...); metadata properties are the extra fields organised by production phase (logline, backstory, appearance, ...). Some are read-only (computed for you) or human-only (never sent to the AI). Ordinary Animator says "property" everywhere -- never "attribute". See Entities.
Recipe -- how a media purpose is made: an ordered series of operations, each one's output feeding the next ("make a still, then animate it"). A recipe names the models it uses, says which shots it is for ("shots with several speakers") and what it needs (an end frame, a recorded performance). You add recipes to a visual style as recommendations. See Models, Workflows & Recipes.
Recipe panel -- the panel beside a gallery that shows its recipe: the steps, each one an operation you fill in and run, with Run all and Choose another recipe. See The Recipe panel.
Recommendation -- one numbered entry in a visual style's list for a media purpose: a recipe, with an art style chosen for each model it uses. The list is read from the top, and the first recommendation that is complete, is for this shot and whose needs exist is used. You add, remove and reorder them on the Visual styles page; deciding an experiment puts its winner first. If none fits, the shot says No recommended recipe fits this shot -- Ordinary Animator never picks a model for you. See Visual styles.
Rough clip -- the clip you have chosen for a shot with Set as rough clip: the one the Sequencer plays until the shot has a finished clip. See the shot's picture.
Scene -- one continuous piece of the story in one place and time, from the screenplay's scene heading. An episode is made of scenes, and a scene is made of shots. A scene can be given a visual style, and every shot in it runs that way. See Scenes.
Screenplay -- the technically precise name for the Script: the specific kind of script used for film and animation (scene headings, action lines, dialogue), as distinct from a stage play or other script forms. Ordinary Animator's scripts are screenplays, written in Fountain.
Script -- the text the human writes, and the source the whole production is driven from: scene breakdown, shots, dialogue, and audio all flow from it. It is not just a view -- it is the authored content. You write it in Fountain format (in any editor, e.g. Google Docs) and bring it in. The technically precise term for it is screenplay. See the Script reference.
Sequencer -- where an episode is cut: the scenes and shots as strips, with tracks of dialogue, music and effects beneath, which you can play and edit. It plays each shot's finished clip, else its rough clip, else its storyboard. See Episodes.
Shot -- the most granular unit of production: one image or video clip. (The old term "panel" is gone -- it is always a Shot.)
Skill -- a written procedure the Assistant loads when your request calls for it: reviewing a still, covering a scene, casting a voice, explaining a field. Skills come with Ordinary Animator, several can apply to one question, and you never have to name one. The project's Skills view lists them all. (Replaces Agent.)
Sound -- dialogue, music and sound effects. Sound has no look, so it is not under a visual style: it has one list of recommendations for the whole show, at the top of the Visual styles page, with no art styles. See Visual styles.
Starter style -- a ready-made visual style that ships with Ordinary Animator ("Anime with MiniMax H3"), its recommendations and art styles already filled in. You read it, then Add to project to get your own copy to change. See Visual styles.
Step -- one piece of work within a phase of the Episode Plan ("Pick the recipes", "Plan the shots"), made of one or more criteria. A step is done when all its criteria are approved.
Storyboard -- a quick preview image for a shot; the place to get blocking (staging) right cheaply before generating a finished clip. Also the level a shot's picture falls back to when it has no finished clip -- set from the shot's gallery, and not restricted to stills. See the shot's picture.
Take -- one version of a line of dialogue: generated, recorded or voice-changed. You choose the take to use (Set as preferred); the newest is used until you do. A shot's chosen clip is not a take -- see Rough clip.
Timeline -- the project view that lays your series out in time: episodes across, tracks down, arcs drawn over the episodes they cover. (Formerly "Arcs".) See the Timeline reference.
Track -- a row on the Timeline, holding the arcs of one thread (a season, a character, a relationship). The Sequencer has tracks too -- rows of picture and sound. Both are rows of things laid out over time. (The old term "swimlane" is gone -- it is always a Track.)
View -- one of the pages of a single entity (Properties, Media, Sequencer, ...), picked from the sidebar under the level it belongs to. (Formerly "modes".) See Navigation.
Visual style -- what a scene means visually ("Normal", "Dream scene", "Flashback"). It is yours, and it belongs to the project. For each media purpose it holds a list of recommendations: recipes in order, the first that fits the shot being used. A scene, a shot or a look can be put in one: a shot's style is its own, else its scene's. A visual style never chooses a model and carries no prompt wording; the wording comes from the art style chosen for each model. See Visual styles.
Voice changer -- turns a recorded performance into a character's voice, keeping the original timing. An alternative to text-to-speech when you want a real performance.
Workflow -- a ComfyUI node graph that generates media. Every workflow is ready to run as it stands. Its App Inputs are the parameters Ordinary Animator fills in, and they are the boxes you see when you choose it as an operation. Workflows are deliberately small and single-purpose, so you can see each intermediate result and re-run just the stage you want.
Workflow library -- a collection of workflows you can copy into a project: the platform catalog, your own project library, your local ComfyUI, or Comfy Cloud. Copying one in is what makes it available as an operation. A workflow may be tuned for where it runs -- a version for Comfy Cloud and a version for your own machine, the latter perhaps using a quantized model so it fits a smaller GPU -- and Ordinary Animator picks the right one for you. See ComfyUI Integration.