Workflow guide

How LizzFlix works

Use this page as the practical map of the studio: what each area is for, what gets sent to providers, and how to move from assets to scenes, episodes, and a finished montage.

Open studio

1. Series library

Start with a series, then keep building inside it

A series is the container for your universe. It holds the character bible, location bible, episodes, scenes, and montage output for one project.

Create a series from the library, then open it right away.
Use the small cards on each series tile to jump directly into episodes, characters, or locations.
Think of the library as your production dashboard, not a storyboard editor.

2. Character and location assets

Build reusable approved assets first

Characters and locations work best when you prepare reusable assets before composing scenes. Approved assets become the reliable material for later start frames and previews.

Upload adds a file to the gallery as-is.
Generate from prompt creates a new asset from text only.
Use image as reference sends the uploaded image to the model as a temporary reference for this one generation.

3. Labels and tags

Use labels for humans and tags for filtering

Asset labels and tags do different jobs, and keeping them clean makes the whole studio easier to use.

Asset label: the human-readable name you see in the gallery, such as "hero close-up" or "cemetery wide establishing".
Tags: searchable descriptors like "smile", "night", "profile", or "red coat".
Labels help you recognize an asset quickly. Tags help you sort, filter, and reuse it later.

4. What goes to providers

Only the active generation input is sent out

LizzFlix does not blindly dump the whole workspace into providers. It sends the fields needed for the specific action you run.

Image generation sends the current prompt, selected model, and any chosen reference images.
Voice preview or TTS sends the chosen voice, model, language, and the line you want spoken.
Scene preview sends the approved start frame, selected video model, and the scene prompt for motion.

5. Voice and dialogue

Save a default voice, then build reusable lines

Each character should have one default voice. You can preview a provider voice, save it as the default, generate reusable voice lines, or create a custom voice from uploaded or recorded speech samples.

A custom voice should always keep a playable example sample so you know what it sounds like.
Voice lines are reusable takes that can be dropped into dialogue timelines later.
Changing the default voice should not delete older usable voices; it just changes which one new TTS defaults to.

6. Compose a scene

A scene is cast + location + prompts + timing

Inside an episode, each scene needs a location asset, character placements, action prompts, and a clip duration. Once those are ready, you can compose a start frame and then render a preview.

Pick the approved location asset and approved character assets first.
Use the placement prompt to describe where the cast stands and how they are framed.
Use the action prompt to describe motion, expression, and what happens during the shot.

7. Add dialogue to a preview

Dialogue is placed on the preview timeline, not on the draft script alone

The dialogue panel lets you prepare lines and audio sources, but the actual clip mix happens on the preview. That is where you choose which takes to use, when they start, and how they fit the duration of the video.

Generate or upload voice lines first if you want reusable takes.
Open a preview, add lines to its timeline, and set exact start times.
Applying dialogue creates a new mixed preview clip with the selected audio layout.

8. Episode flow

An episode becomes a queue of finished scene clips

Episodes are built scene by scene. You usually save the scene, compose the start frame, generate previews, choose the one you want, and promote it to the final slot for that scene.

Previews are for exploration and dialogue timing.
The final slot is the approved scene clip that should feed montage.
A scene can have many previews, but only one final should matter for assembly.

9. Montage

Montage assembles the final episode

Montage is where approved final scene clips are ordered into one episode export. If a scene already had dialogue applied before promotion, that dialogue is already inside the final clip you bring into montage.

Sync finals into montage after scenes are approved.
Reorder clips, add montage notes, and prepare assembly.
Export the full episode only after the montage reflects the final scene order.

10. Background jobs

Generation should not trap you on one page

Most heavy actions run as background jobs on Cloudflare. The page can show a placeholder card and a status box while the job runs, but you should still be able to leave and come back later.

If you stay on the page, the UI will update when the job finishes.
If you leave, the result should still appear when you return.
Use the visible status boxes as progress hints, not as a sign you must wait on one screen.