What is AI Studio?
AI Studio is Pictory's workspace for creating and editing AI-generated images and creating AI-generated video clips. It uses a chat interface, so the prompts and results for each project remain available in the project's history.
To open AI Studio, open a project and select Visuals, then AI Studio.
Create an image or video
Select Images or Videos above the prompt box.
Write a prompt. If your project has scene text, you can also select Get a prompt for this scene to create a starting prompt.
Optional: attach one or more reference images. The number and file types allowed depend on the selected model.
Open Settings and choose the model, aspect ratio, and other generation options.
Review the estimated AI credit cost below the prompt box.
Select Generate.
AI Studio remembers your last-used generation settings for each project, including the model, aspect ratio, duration, audio choice, number of variants, Brand Kit, and training preset. A direct action such as Edit with AI, Animate with AI, or Extend can change the mode and settings needed for that action.
Create and edit images
Select Images to create a new image from a prompt. In Settings, you can choose an image model, visual style, aspect ratio, Brand Kit, training preset, and the number of variants.
To edit an image, use one of these methods:
Select the attachment button and upload an image.
Paste an image into the prompt box with Ctrl+V on Windows or Command+V on macOS.
Select Edit with AI from an existing image's menu.
Use a generated image as a reference from its result menu.
Describe the change you want, then generate the result. AI Studio selects an editing-capable model when you enter from an image-editing action. If a model cannot change the source image's aspect ratio, AI Studio locks the aspect ratio to the source image.
Available image models
Flux
Nano Banana Lite
Seedream
Nano Banana
Nano Banana Pro
Nova Canvas
Model availability and supported settings can vary. AI Studio updates the controls when you change models. Amazon Titan has been retired and no longer appears in the model picker, but results created with Titan remain visible in older chats.
Guide a result with reference images
Reference images help a model retain a person, product, setting, composition, or visual style.
Select the attachment button and add your images. You can select several files at once when the model supports multiple references.
In the prompt, type @ or select the @ button.
Choose a reference to insert a token such as @image1.
Describe how the model should use each referenced image.
For example: Place @image1 in the office shown in @image2 and preserve the product colors.
Pixverse can use up to four reference images. Other models may allow fewer references or different file types. If you switch to a model that cannot use the staged references, AI Studio removes the incompatible items, explains why, and offers Undo.
Reference numbers identify the attached files. If you remove a reference, AI Studio does not reassign its number to another image. Remove any matching @imageN token from the prompt before generating.
To suggest a transition between two images, attach both images, mention both in the prompt, and describe the transition. AI Studio does not have a separate first-frame or last-frame control.
Create videos with optional audio
Select Videos.
Write a prompt or create one from the current scene.
Optional: attach reference images and mention them in the prompt.
Choose the video model, duration, aspect ratio, Brand Kit, training preset, and number of variants.
Use the sound switch to turn generated audio on or off when the model supports this choice.
Review the estimated AI credit cost, then select Generate.
Some models always create audio. For example, Omni Flash keeps audio on. Other models may not support generated audio. AI Studio locks or disables the sound switch to match the selected model.
Available video models
Pixverse: supports reference images, video extension, generated audio, durations up to 15 seconds, and a wide range of aspect ratios.
Veo Fast: creates cinematic clips with shorter generation times.
Veo: creates higher-quality cinematic clips.
Omni Flash: creates 10-second videos with synchronized audio.
Pictory may route a Pixverse request to the model version that supports the selected features. The model picker therefore shows Pixverse without a version number. The credit estimate below the prompt reflects the version and options that Pictory will use.
Extend an existing video
Use Extend to continue a generated clip or an uploaded video with a new prompt.
Select Extend above the prompt box. You can also select Extend from a generated video's menu.
Upload or select one source video. AI Studio accepts only one source clip for an extension.
Describe what should happen next.
Choose an available model and extension duration.
Review the estimated AI credit cost, then select Generate.
The result is a single clip containing the source video and its continuation. Available models and duration controls depend on provider support. Veo extensions add a fixed seven-second continuation, so AI Studio locks the duration when you use a supported Veo extension model.
Generate up to four variants
Variants let you compare several results from the same prompt without submitting each request separately.
Open Settings.
Under Variants, choose a value from 1 to 4.
Generate your image or video.
AI Studio displays multiple results in a grid. Each variant is a separate generation and uses AI credits. If one variant fails, the successful variants remain available, and Pictory charges only for the variants that were delivered.
Use your Brand Kit
To guide a generation with your brand:
Open Settings.
Select Reference brand kit.
Choose a Brand Kit from the project workspace.
AI Studio uses the selected brand's name and colors as generation guidance. Select None to generate without Brand Kit guidance.
Use a training preset
Training presets add ready-made guidance for common training content. Open Settings, select Training preset, then choose:
Sales training
Product training
Soft skills training
Select None to use only your prompt and other settings.
Reuse a prompt or generate again
Reuse restores the prompt, model, references, aspect ratio, duration, audio, variants, Brand Kit, and training preset in the composer. You can edit the request before sending it.
Generate again repeats the original request with the same available settings.
If an older result used a retired model, AI Studio selects a current model from the same image or video category.
Use generated results in your project
Open a result's menu to use the generated asset. Depending on the asset and your current workflow, you can edit or extend it, use it as a reference, add it to My Uploads, or apply it to the project.
AI Studio does not automatically save every result to My Uploads. Select Add to My Uploads for each result you want to keep in your media library.
Understand AI credit charges
AI Studio shows an estimated credit cost below the prompt box before you generate. The estimate changes when you select a different model, duration, audio option, or number of variants.
Video charges depend on the model, delivered duration, and audio option.
Each delivered variant is charged separately.
Failed variants are not charged.
If no result is delivered, Pictory releases the reserved credits.
For current rates and information about purchasing more credits, see Understanding AI Credits.
Troubleshooting
The Generate button is disabled: wait for attached files to finish uploading, add a source video in Extend mode, remove an unmatched @imageN token, or choose a model that supports the requested action.
A setting changed after I selected another model: models support different durations, aspect ratios, audio options, file formats, and reference limits. AI Studio adjusts incompatible settings and shows a notice.
I cannot edit an image: choose an image model that supports editing. AI Studio normally selects one when you enter through Edit with AI.
I cannot attach a video: switch to Extend. Videos are source clips for extension, not image references.
I cannot switch modes during generation: wait for the current image or video generation to finish or stop it before changing modes.