Viz 360

The AI Director's Workspace. Plan your film, preview it, and generate only the shots you approve.

Comment

Inspiration

Today's AI video workflow is broken. Creators jump between chat tools, image generators, video generators, editors, and timelines. Context gets lost between every step, iterations get expensive, and many generated clips never make it into the final edit.

Viz 360 was built to solve that waste. Instead of generating videos first and fixing them later, Viz 360 gives creators an AI-first pre-production workspace where the story, shots, characters, environments, and timeline stay connected from the beginning.

What it does

Viz 360 transforms a screenplay into a complete production plan.

A creator starts with a project idea or script. From there, Viz 360 helps:

  • generate or edit the screenplay;
  • break the story into ordered scenes;
  • break scenes into shots;
  • generate progressive keyframes and storyboard prompts for each shot;
  • organize everything on a timeline;
  • play the story back as a lightweight animatic or story preview.

The key idea is simple: validate the film before spending heavily on final generation.

Instead of rushing into expensive video generation, Viz 360 creates a playable story preview using lightweight assets such as text-first boards, prompts, references, uploaded images, and temporary placeholders. That lets creators review pacing, shot order, continuity, and intent before committing to final renders.

Why it's different

Most AI video tools optimize generation.

Viz 360 optimizes decisions.

The goal is not "how can we generate more videos?" The goal is "how can we avoid generating the wrong videos in the first place?"

By preserving project-wide creative context, Viz 360 helps creators reduce wasted generations while improving consistency across characters, environments, and story beats.

Core workflow

Idea -> Script -> Scenes -> Shots -> Keyframes -> Story Preview -> Final Video

Viz 360 acts as the production layer between concept and render.

What is working now

The current product already includes a working local demo flow with:

  • project creation and a multi-project dashboard;
  • a workspace with View, Script, Characters, and Environments tabs;
  • screenplay generation and editing;
  • scene generation from saved script context;
  • shot generation from scene context;
  • progressive keyframe generation and storyboard prompt generation;
  • text-first preview playback at project, scene, and shot levels;
  • timeline-style scene and shot editors;
  • CRUD for projects, scenes, shots, keyframes, and layers;
  • SQLite-backed local persistence;
  • provider support for OpenAI, Gemini, and local Ollama-based testing.

The product is intentionally honest where full media generation is still incomplete. Instead of faking finished visuals, Viz 360 shows the exact prompts, script coverage, context, and keyframe intent directly in the preview and timeline. Creators can also upload or paste their own images into shots and use those as visual steps toward an animatic.

How AI is used

Viz 360 uses AI throughout the planning pipeline.

It can:

  • turn a project idea into a screenplay;
  • analyze screenplay and reference material to set up creative direction;
  • break a story into scenes;
  • break scenes into shots;
  • create progressive keyframes that describe how action evolves over time.

Because the app maintains project-wide context, each downstream generation can include the relevant script, characters, environments, scene purpose, shot intent, and prior keyframes.

How we built it

Viz 360 is built as a Next.js application with React, TypeScript, SQLite, and server-side AI provider adapters. The current implementation includes structured server APIs, local demo authentication, timeline-oriented editing components, persisted project hierarchies, and browser-tested core flows.

The architecture is designed so creators can use lower-cost local models while planning, then move to higher-quality cloud models when they are ready.

Challenges we ran into

The hardest challenge was not generating AI content. It was preserving creative context across an entire film.

Characters, environments, scene intent, shot purpose, timing, and camera direction all need to stay connected. The project needed to be more than another prompt box. It needed a production structure creators could actually use.

Another challenge was staying honest in the product experience. Where final image or video generation is not fully connected yet, Viz 360 avoids fake outputs and instead shows a clear, usable planning layer that creators can review and edit today.

Accomplishments that we're proud of

  • Built an AI-first production workspace from scratch.
  • Unified screenplay, scene, shot, and keyframe planning in one app.
  • Designed a timeline-first workflow for story preview and review.
  • Created a playable text-first animatic workflow instead of isolated prompt steps.
  • Preserved project-wide context across the creative pipeline.
  • Reduced unnecessary AI generation by helping creators validate decisions earlier.

What we learned

The future of AI filmmaking is not just better prompting. It is better planning.

Creators need a director's workspace, not only a generation box. Context continuity matters as much as model quality.

What's next for Viz 360

  • connect production-grade OpenAI image and video generation;
  • add stronger character and environment locking;
  • add persistent media storage and background render jobs;
  • support collaboration, review, and approvals;
  • export finished animatics and final videos;
  • add a hosted judge-friendly deployment.

How Codex and GPT-5.6 were used

Codex and GPT-5.6 were used throughout the project as engineering and product-design partners. They helped shape the product architecture, define the production hierarchy, build reusable React and API flows, refine the timeline workflow, fix bugs, validate builds, and improve the overall UX and product story.

They were especially valuable in turning a broad idea about AI filmmaking into a structured product with a clear workflow: project to script, script to scenes, scenes to shots, shots to keyframes, and keyframes to story preview.

Built With

Share this project:

Updates