Generate media
Create images, video, speech, music, and sound effects with the studio tools.
Besides asking Rolit in plain language, you can start a specific kind of generation from the creation tools menu in the chat box. Selecting a tool shows its settings above the message box. Generated files are saved as artifacts in the project's Studio, and long-running generations appear under Jobs.
Each tool uses the model you chose in Choose a model. Costs are charged in credits.
Tools
| Tool | What it makes |
|---|---|
| Script | A script tailored to your style. |
| Research | A research brief on a topic. |
| Image | Images from a prompt. |
| Thumbnail | Click-worthy thumbnails. |
| Audio | Voiceover and speech. |
| Music | Original music tracks and sound effects. |
| Video | Short-form video. |
| Captions (Transcribe) | A transcript from audio or video. |
| Hashtags | Platform-optimized hashtag sets. |
| Paper Animator | Paper-cut and text animations. |
The rest of this page covers the media tools in detail.
Images
Set these options before you send your prompt:
| Setting | Options |
|---|---|
| Size | 1024x1024, 1024x1792, 1792x1024 |
| Images | 1, 2, 3, or 4 |
| Quality | standard, hd |
| Style | natural, vivid |
You can also pick a reference image from the project or upload one, and Rolit uses it as the starting point.
Video
Video settings include:
| Setting | Options |
|---|---|
| Duration | 4 to 15 seconds |
| Aspect ratio | 16:9, 9:16, 1:1 |
| Resolution | 720p, 1080p, 4k |
| FPS | 24 or 48 |
| Reference image | Starting frame for image-to-video. |
| Reference audio | Audio to accompany the video. |
| Disable safety filter, Disable prompt upsampling | Toggles for supported models. |
Which settings appear depends on the selected model. For example, avatar video offers 720p and 1080p only.
Avatar video
Turn on avatar video mode to make a portrait speak:
- Select a portrait image (required).
- Give the avatar something to say with a Voice Script, or select Speech Audio, which overrides the script.
- Optionally set the Voice, Voice Language, a Voice Prompt for style, tone, pacing, or emotion, and a Video Prompt.
Voice languages include English (US and UK), Spanish, French, German, Italian, Portuguese (Brazil), Japanese, Korean, and Hindi.
Audio and speech
Choose a Model, an optional Voice, and a Format: mp3, wav, opus, flac, or pcm. Models that use Gemini text-to-speech offer a list of named voices.
Music and sound effects
The Music tool opens a Song Canvas:
- Describe the song, or write your own lyrics. Use markers such as
[Verse 1]for structure. - Set Tempo from 30 to 300 BPM, or leave it on AUTO.
- Set Track Length from 30 to 300 seconds.
- Choose Vocals or Instrumental. Instrumental ignores any lyrics.
- Choose the Engine (music model).
The same tool also makes short sound effects. In the project's Music studio, a Tracks / Sound FX switch toggles between the two.
Track progress
Image, video, and music requests run as background jobs with a status of Queued, In progress, Completed, or Failed. Open Jobs in the project to follow them. Failed jobs show the error message.