Skip to content

Mastering Lorvara

The Lorvara power user workflow is typically:

Step What it is
Project An organized set of prompts that make up one piece of work. It attaches a brief; several projects can use the same one.
Brief The world your production takes place in: look, intent, cast, locations, props, glossary. Written once, read by everything after it.
Scenes The grouping of shots that share characters, props, environment plates and more.
Shot list The work order for one shoot. Related shots make up a scene.
Key Prompts Each shot becomes a prompt plus its model, parameters and references. That is the unit you keep and re-run.
Generate The model renders it. The result lands in your library, attached to the prompt that made it.

Briefs provide context within a project. Once you create a Project and attach a Brief, Lorvara parses your prompts based on the job they were assigned to do. For example, some prompts will be responsible for creating reference images, and some prompts will be responsible for final output.

Scenes organize shots together, group a series of shots together as prompts that work together to produce media that will be edited together from the clips produced. Scenes typically share references such as plates for a specific environment, a cast of characters in the scenes, and elements within that scene.

Shots become prompts within a project that are typically built for video generation within the context of a particular scene within a brief. Shots generally take in reference media from supporting prompts in a project. Working in the Studio, these prompts are dialed in through iteration in the Studio. Lorvara does much of the heavy lifting by allowing you to manage references in the brief that will, in turn, appear in your shot.

Use your LLM to create the brief. The more time you invest in getting the brief as close to your final production, the more time Lorvara will save you during production. Use our template or downloadable skill to assist you with this process until you’re ready to import into Lorvara. Get more familiar with brief structure.

How Lorvara reads your brief. Once ready, upload or copy & paste your brief into Lorvara after creating a new brief within a project (or to attach to a project later). Know that when Lorvara parses your brief, it will then propose a series of blocks for you to approve. This includes the Look, Intent, Cast, Locations, Props and Glossary. Note that Lorvara will guess when there are gaps. You don’t need to accept these guesses, and you can delete or edit anything before the brief is confirmed in the system.

Review the Shot List carefully. Once the brief blocks are confirmed, Lorvara will either create or derive a shot list based on your input. Again, to save time down the road, review each scene and shot carefully. Look at the supporting reference image prompts to make sure you have the key elements you need for final output. You can still add shots, scenes and references later, but dialing them in at this point saves time and frustration.

Knowing how Batch Prompts work. Once you confirm your Scenes & Shots, Lorvara will begin parsing out all the prompts needed to create the media for your project. This will take several minutes. Note that the number of prompts that can be batched together is 40. So complex creations may require multiple projects and briefs. Larger projects will require multiple Projects and Briefs inside Lorvara.

Start with REFS and move to final prompts. Once your list of prompts lands in the Project, it’s time to begin working in the studio. Start with the Reference prompts to build up your library of reference images for the final shot prompts. As you generate, media is transferred to the Library, a home for media organized by prompt. As you produce reference images, go back and add them to the Brief. You will then recompose prompts so they now carry the proper reference images.

Recomposing prompts matters. As you add reference images into the Brief, you can recompose at the point of the reference, or you can recompose in the Scenes & Shots. It’s more economical to create and add all your references at once, avoid recomposing individually, and instead, recompose your shots in the Scenes & Shots page. That way, all the references can be added at once versus adding them one at a time.

Here are a few of the advanced features that get you to your goals faster:

Feature When to use
Prompt Agent Inside the Studio, the Prompt Agent, powered by Claude, will help create your prompts just like your LLM, but with more intelligent context.
Brief Agent More than a handful of prompts belong to one piece of work and you have started scrolling to find them.
A Skill When looking for the same kind of shot, you can use skills to guide the prompt agent to replicate a specific style.
Skill Agent You can use built-in skills, import your own, or use the Skill Agent to guide you through the process of building your own skill.
Batch generation When parsing a shot list from a brief or just reparsing a specific shot once you’ve added a set of reference images, the parsing work can be viewed in a project’s batch prompts.
Gallery You have the option to make the cover image of any prompt a part of a gallery where you can view media across teams and projects. Great for seeing a grid of completed work.

The order matters less than the trigger. Most people arrive at Skills first, because it is the one that changes the picture rather than the paperwork.