Help Center

Projects & Workflow

A structured production pipeline from concept to finished work.

Production hierarchy

Every project follows a clear structure: Project → Story → Scene → Shot → Asset. This mirrors the pipeline used by professional studios, keeping your work organized from initial concept to final delivery.

  • Project — The top-level container. A project represents a complete production (a short film, an ad campaign, a game asset library, etc.).
  • Story — A narrative arc within your project. A project can have multiple stories for different episodes, chapters, or campaign variations.
  • Scene — A single scene within a story. Scenes group shots that share a location, time, or narrative beat.
  • Shot — An individual shot within a scene. Each shot can hold multiple asset versions — generate, compare, and approve the best take.
  • Asset — A generated or uploaded file (image, video, audio, 3D, avatar) attached to a shot. Assets are versioned automatically.

Creating a project

From the dashboard, click “New Project” and give it a name and description. Projects move through five production stages:

  1. Development — Brainstorming, concept development, and initial research.
  2. Pre-production — Script writing, scene planning, shot lists, and asset design.
  3. Production — Active asset generation, iteration, and refinement.
  4. Post-production — Final polish, review, and approval of all assets.
  5. Complete — Project finished and archived.

You can advance the project stage at any time from the project settings. Stages help you and your team track where the production stands at a glance.

Concepts & the AI assistant

Before generating assets, use the concept workspace to develop your creative vision with the help of the AI assistant:

  • Concept briefs — Describe a rough idea and the AI assistant helps you refine it into a structured creative brief with visual direction, tone, and reference notes.
  • Story writing — Develop narratives, scripts, dialogue, and lyrics with AI-assisted writing.
  • Scene breakdowns — Automatically break a story into individual scenes with descriptions, mood notes, and suggested shot lists.
  • Shot planning — For each scene, generate detailed shot descriptions that can be used directly as generation prompts.
  • Language setting — Set your content language (Tamil, Hindi, Spanish, etc.) once in the concept. All AI writing — dialogue, lyrics, scene descriptions — generates in your chosen language automatically.

Concepts are saved alongside your project so you always have a record of the creative decisions that led to your final assets.

Stories, scenes & shots

Inside a project, build out your narrative hierarchy:

  • Add a story — Create one or more stories to outline your narrative arcs. Each story can have its own synopsis, script, and scene list.
  • Break into scenes — Divide each story into scenes. A scene groups all the shots that share a location, time of day, or narrative beat.
  • Plan your shots — Within each scene, create individual shots. Each shot has a description, camera angle notes, and a place to attach generated assets. Shot durations are planned to match your scene's time budget using values supported by video models (4, 5, 6, 8, or 10 seconds per shot).

This hierarchy is optional — for simple projects, you can skip straight to generating assets. But for larger productions, it keeps everything organized and makes it easy to track progress across dozens of assets.

Runtime tracking

Rhyora tracks your project's total runtime automatically. As scenes and shots are planned, their durations are aggregated to show:

  • Total estimated runtime — shown in the project header as a badge (e.g., “Est. 5m 30s”).
  • Format target validation — if your project uses a production format (Short Film, Feature Film, Commercial, etc.), the system checks whether your planned runtime fits the expected range and warns you if it's over or under.
  • Production progress — a time-based progress bar shows how much of your estimated runtime has been produced (shots marked as “Ready”).

Shot durations use values supported by video generation models (4, 5, 6, 8, or 10 seconds). For longer scenes, the AI generates more shots rather than longer individual shots. This ensures every planned shot can be generated without duration mismatches.

Asset versioning

Every generated asset is automatically versioned with a unique ID following professional studio naming conventions:

IMG-20260203-a3f8-001_v001.png

The ID encodes the asset type, date, project reference, sequence number, and version. When you regenerate or edit an asset, the version number increments automatically. You can:

  • Compare versions side by side to evaluate quality.
  • Promote any version to the active/approved version.
  • Track the full generation history of each shot.
  • Download any version at any time — nothing is overwritten.

Review & approval

Shots follow a status flow that tracks their readiness:

  1. Empty — No assets generated yet.
  2. WIP — Assets generated, still iterating.
  3. Review — Ready for review by the project owner or team lead.
  4. Ready — Approved and locked for delivery.

Individual assets also have a status: Draft (default) or Approved. Only approved image frames can be used as input for video generation — this quality gate ensures you only invest video credits in frames you're satisfied with.

Produce shots from the storyboard

Click any empty shot card directly on the storyboard to start producing it. A guided modal walks you through each step without leaving the storyboard:

  1. Generate frame — An image keyframe is created using your project's preferred model. A progress indicator shows the current step.
  2. Review your frame — The frame appears in a preview. You can approve it, try again, or suggest changes (describe what to fix and the AI regenerates with your feedback).
  3. Produce video — Once you approve the frame, video generation starts automatically. The approved frame stays visible so you know exactly what's being animated.
  4. Review your video — Watch the video right in the modal. Approve to finalize the shot, or try again with the same frame.

The shot card on the storyboard shows a live producing indicator while generation is running, so you always know which shots are in progress even if you close the modal.

Suggest changes

During frame or video review, click Suggest Changes to describe what you'd like to adjust — for example, “make the lighting warmer” or “zoom in on the character.” The AI regenerates with your feedback applied to the original prompt. This lets you refine results without manually editing prompts.

Rhyora Prism

The Rhyora Prism can produce your entire storyboard with minimal intervention. Choose how much control you want:

  • Manual — The agent works through your storyboard in phases and pauses at key checkpoints for your approval. You review proof shots, approve image batches, and confirm before video production begins. Recommended for most projects.
  • Auto — The agent runs the full pipeline on its own. Set a credit budget and a quality threshold, then let it produce every shot. You can watch progress in real time, pause at any point, or provide feedback while it works.

Choose your agent mode from Project Settings. Both modes include these built-in safeguards:

  • Prompt optimization — Before the first image attempt, AI enhances your shot prompt with technical photography terms, lighting details, and composition guidance for better first-attempt results.
  • Quality gate — After each image is generated, AI vision evaluates it against your prompt for composition, technical quality, and mood. Images that meet your quality threshold are auto-approved. Images below the threshold are automatically retried with an improved prompt.
  • Smart retries — If a generation fails due to a server issue, the agent retries the same prompt (since the problem was infrastructure, not content). If the model rejects the prompt, the agent simplifies it. Each shot is limited to 3 retries to prevent runaway credit usage.
  • Credit budget — In Auto mode, the agent stops when the budget you set is reached. In Manual mode, you approve each phase before credits are spent.

The Activity Log at the bottom of the production panel shows every action the agent takes — generations, quality evaluations, approvals, retries, and budget usage — so you always know what happened and why.

Produce All

To produce multiple shots at once, click the Produce All button in the storyboard toolbar. It starts generating frames for all empty shots that have prompts, up to your plan's parallel task limit:

  • The button shows how many shots will be produced and how many parallel slots are available.
  • Each shot card updates with a producing indicator as its generation starts.
  • If you reach the concurrency limit, remaining shots start automatically as slots free up.

For advanced control — choosing specific models, linking references, or adjusting individual parameters — open a shot in the full production view via the hover popover's Open in Production link.

Post-production

The Post-Production tab consolidates all post-production work with sub-tabs:

  • Dialogue — Write and refine timed dialogue for each shot with dialect support and AI refinement.
  • Lyrics — Compose song lyrics with genre-aware AI generation and chat-based refinement.
  • Audio — Voice-over, music, and SFX generation (coming soon at the project level).
  • Songs — Full song composition workspace (coming soon).

Project settings

The Settings tab (at the far right of the phase tabs) lets you configure default generation parameters:

  • Image Model — Default model for image generation.
  • Video Model — Default model for video generation.
  • Quality — Low, Medium, or High — mapped to the closest option each model supports.
  • Agent Mode — Choose Manual (you drive, approve each step) or Auto (runs the full pipeline within your budget).
  • Quality Threshold — Minimum quality score (1-10) for auto-approving generated images. Lower values approve more freely; higher values demand closer prompt adherence.
  • Credit Budget — Maximum credits the Auto agent can spend per session.

Changes auto-save immediately and apply to all shots using the Produce Shot pipeline.

Downloading & exporting

Every asset can be downloaded in its original format and resolution. From the asset detail view or the project asset list, click the download button to save the file locally. Supported export formats:

  • Images — PNG, JPG
  • Video — MP4
  • Audio & Music — MP3, WAV
  • 3D — GLB, OBJ

Frequently asked questions

Do I need a project to generate content?

No. You can generate assets directly from any studio without a project. Projects are useful when you want to organize larger productions with multiple related assets.

Can I move assets between projects?

Assets belong to the project they were generated in. You can download and re-upload assets, but direct transfer between projects is not currently supported.

What does "approve" mean for an asset?

Approving an asset marks it as production-ready. Only approved image frames can be used as input for video generation. This quality gate prevents wasted credits on videos from frames you'd reject.

How does versioning work?

Every generation and edit creates a new version automatically. Versions are named with studio-style IDs (e.g. IMG-20260203-a3f8-001_v001). You can compare versions side by side and promote any version to approved.

How does runtime tracking work?

When scenes and shots are generated, the AI assigns durations that fit your production format (e.g. Short Film targets 5-20 minutes). Shot durations use values supported by video models (4, 5, 6, 8, or 10 seconds). The project header shows total estimated runtime, format target validation, and a time-based production progress bar.

Can I produce multiple shots at the same time?

Yes. Click "Produce All" in the storyboard toolbar to start generating frames for all empty shots that have prompts. The number of shots produced at once depends on your plan's parallel task limit. Each shot card shows a live producing indicator while generation is running.

What happens if I close the production modal mid-generation?

The generation continues in the background. The shot card on the storyboard will show a producing indicator. When you return, shots with new assets will be ready for review.

What is the Rhyora Prism?

The Rhyora Prism produces your storyboard shots automatically. In Manual mode, it pauses at checkpoints for your approval. In Auto mode, it runs the full pipeline within a credit budget you set. Both modes include prompt optimization, quality evaluation, and smart retries.

How does the quality gate work?

After generating an image, AI vision scores it on prompt accuracy, composition, technical quality, and mood (scale of 1-10). Images above your quality threshold are approved automatically. Images below the threshold are retried with an improved prompt. After two quality retries, the image is approved to keep production moving.

Can the agent run out of control and burn credits?

No. Multiple safeguards prevent this: a credit budget cap, a per-shot retry limit of 3, a quality retry limit of 2, and a circuit breaker that pauses after 5 consecutive failures. You can also pause or stop the agent at any time.

Ready to start your production?

Create your first project and build from concept to finished asset.

Start Free