The last chapter argued that the model is one instrument in a large orchestra. This chapter introduces the orchestra. It is a walk through the whole board, phase by phase, naming the companies doing the work in each box, so the abstract map becomes a place you can actually navigate. Read it not as a directory to memorize but as a tour: what each part of the field is for, why it exists as its own business, and who is building it in 2026. The full company-by-company index lives in the ecosystem appendix; this is the guided version.
The field starts before anything is generated. The development phase is the writing room and the planning office: the tools that turn an idea into a script, a script into shots, and shots into a plan you can actually shoot. It is the least glamorous phase and the one most people skip, which is exactly why it is where a lot of the quality is quietly won or lost.
Three jobs live here. The first is story itself. Most AI films still begin in a text model: writers draft and structure with general assistants like Claude, ChatGPT, and Gemini, and fiction-specific tools like Sudowrite add story-aware drafting on top. The second job is turning that script into pictures you can direct from. Storyboarding tools like LTX Studio, Storyboarder.ai, and SAGA generate boards and animatics straight from a script, so a director can see the film before a single expensive generation runs. The third is the unglamorous machinery of production planning, where Filmustage, StudioBinder, and Cuebric handle breakdowns, scheduling, and virtual-production backgrounds, importing the discipline of a real production office into an AI-native pipeline.
The lesson of the development phase is that generation is not the beginning of the work. The people who get good results are the ones who did the writing and the shot design first and then generated against a plan, rather than prompting into the void and hoping.
The Blank Page, Assisted
Story Development is the writing layer of the pipeline, where an idea becomes an outline and an outline becomes a script. It sits before any frame is generated, and in practice it is where AI first entered most creative workflows, because language models were useful for ideation, structure, and drafting long before video models were watchable. The tools here divide roughly into general assistants that happen to be good at narrative and purpose-built writing software aimed at screenwriters and novelists.
The three frontier assistants anchor the category. Claude, Anthropic's model, is the long-form workhorse of the group, strong at holding structure across a full draft; this guide itself was written with it. ChatGPT, OpenAI's assistant, is the most broadly adopted for ideation, outlines, and script drafts, the default that most writers reach for first. Gemini, Google's assistant, carries the same conversational drafting ability and is tied into Google's Gemini media models, which makes it a natural bridge for teams already generating images and video in that ecosystem.
Around the general models sit the specialists. Sudowrite is the fiction-focused option, built for novelists with story-aware drafting features that track characters and continuity rather than treating each prompt in isolation. On the screenwriting side, Arc Studio pairs conventional script software with collaboration and AI assists, and Prescene leans into script analysis alongside writing. A layer below those, Grammarly and Hemingway Editor are line-level editors, one for grammar and tone, the other for flagging dense or passive prose, useful in a writing room but not story engines.
The remaining names, Cursor, Laper, and the rest, are long-tail here. Cursor is an AI code editor that shows up only because technical teams use it to build creative pipelines, not to write scripts. The category's center of gravity is the frontier assistants plus the two or three screenwriting tools; everything else is a supporting utility.
From Script to Shots
Storyboarding is the step where a written script becomes a visual plan: a sequence of frames that describe shots, camera angles, and staging before anything is filmed or generated. In traditional production this is a slow, hand-drawn craft. The AI tools in this category compress it by reading a script or a prompt and generating frames and animatics directly, which lets a small team see a rough version of a sequence in an afternoon rather than a week. The category exists because previsualization is the cheapest place to make expensive decisions.
LTX Studio is the most ambitious entry, an end-to-end system that takes a script and produces storyboards and previs across a whole project, positioning itself less as a single tool than as a pre-production environment. Storyboarder.ai and Reelfy work at the more direct script-to-frames level, generating storyboards from a script or prompt, with Reelfy pushing further into animatics, the timed, moving version of a board. SAGA frames itself as a pre-production workspace built around storyboarding, aimed at teams that want the boards to live alongside the rest of their planning.
Boords is the notable case of an established storyboarding tool that added AI frame generation rather than starting from generation, which gives it a foothold with people who already board professionally. 10Zebra rounds out the group on the shot-planning side. This is a compact category without a long tail of irrelevant names; the six here are more or less the field, and the meaningful distinction between them is how much they try to own, from a single board to the entire previs stage.
The Breakdown and the Board
Production Assistant covers the logistics layer of pre-production: reading a finished script and turning it into the documents a shoot actually runs on. That means script breakdowns, which tag every prop, character, location, and effect a scene requires, plus the scheduling and call sheets that follow from them. This is unglamorous, detail-heavy work that AI is well suited to, because a language model can parse a screenplay and extract structured elements far faster than a human assistant director marking up pages by hand.
StudioBinder is the established production-management incumbent in the group, offering breakdowns, call sheets, and scheduling as a full suite, and it is the reference point most professionals already know. Yamdu plays a similar role as a broader production planning and collaboration platform. Filmustage is the more AI-forward specialist, built specifically around automated script breakdown and scheduling, and RivetAI targets the same core function of analysis, breakdown, and scheduling from the script. Between them these tools define the category: take the script, produce the plan.
Cuebric is the odd and interesting one, because it addresses virtual production rather than paperwork, generating AI backgrounds and 2.5D environments for LED-wall and in-camera work. It belongs here as a production-support tool but solves a different problem than the schedulers. The rest, Drawstory, FilmPro, Studiovity, SHAI Creative, and FinalBit, are long-tail: a mix of previs assistants and pre-production toolkits that overlap heavily with the leaders above without yet distinguishing themselves. The names worth knowing are StudioBinder and Yamdu for management, Filmustage and RivetAI for AI breakdown, and Cuebric for the virtual-production angle.