Video, voice and image nodes
Generate videos and voiceovers, edit images and videos, and add captions inside a workflow.
The five media nodes
These nodes let a workflow turn an idea into a finished, captioned video and publish it, all on its own schedule. Add them from the Add node menu, wire them together, and double-click a node to set it up.
| Node | What it does | Cost |
|---|
| AI Video Generator | Makes a video from a prompt, or animates an image wired into Start Image. | Same price as the Video studio |
| AI Voiceover | Reads text aloud with Kokoro or ElevenLabs voices. | 40 or 100 tokens per 1,000 characters |
| Image Tools | Upscales an image, removes its background, or edits it from a prompt. | 15 tokens, or the edit model's price |
| Video Captions | Listens to the speech and burns subtitles into the video. | 200 tokens per video |
| Video Edit | Trims, merges clips, adds audio, resizes, adds text or a watermark. | Free, up to 30 edits an hour |
Fair pricing. Paid nodes charge what the AI model costs us, converted to tokens (at least 5 tokens). If a node fails, its tokens are refunded automatically.
Example: a narrated, captioned Reel
- AI Text Generator writes a script
- AI Voiceover reads it
- AI Video Generator makes the shot
- Video Edit adds the voiceover
- Video Captions adds subtitles
- Instagram publishes it
- Write the script: Add an AI Text Generator and ask for a 20-second narration. Wire its output into an AI Voiceover node's Text In.
- Make the video: Add an AI Video Generator, pick a model and describe the shot. Pick a duration close to the narration's length.
- Put the voice on the video: Add a Video Edit node set to Add audio. Wire the video into Video In and the voiceover into Extra In.
- Add captions: Wire the Video Edit output into Video Captions. Pick the colour and position.
- Publish: Wire the result into an Instagram, TikTok or YouTube node, then run the workflow or put it on a schedule.
AI Video Generator
Pick any of the models from the Video studio, then set the duration, shape, resolution and sound. Text wired into Prompt In replaces the prompt typed in the node. Wire an image into Start Image to animate it instead of starting from text.
Videos take from one to several minutes. The workflow waits for the video before it continues, and the finished video also appears in your Video studio history.
AI Voiceover
| Model | Best for | Price |
|---|
| Kokoro | Fast, natural English narration at the lowest price | 40 tokens per 1,000 characters |
| ElevenLabs Turbo v2.5 | Expressive voices in 32 languages | 100 tokens per 1,000 characters |
Scripts can be up to 5,000 characters. Use Preview voiceover in the node settings to hear the voice before running the workflow; the preview is charged like a normal run.
Video Captions
Captions are made from the speech in the video, so the video needs someone talking. For AI videos without dialogue, add a voiceover first with AI Voiceover and Video Edit. Captioning takes about a minute.
Video Edit
| Edit | What to wire in | Good to know |
|---|
| Trim | The video | The cut can start up to a second early, at the nearest keyframe. |
| Merge clips | Up to 5 videos | Joined in the order you added the wires. |
| Add audio | The video, plus audio into Extra In | Replace the sound, or keep the original quietly underneath. |
| Resize | The video | 9:16 for Reels, Shorts and TikTok, 1:1 or 4:5 for posts, 16:9 for YouTube. |
| Text overlay | The video | Add a title at the top, middle or bottom. Supports {{ expressions }}. |
| Watermark | The video, plus an optional logo into Extra In | Without a logo, your text is added as a badge in the corner. |
Limits. Each video can be up to 2 minutes and 100 MB. Merge, resize, text and watermark re-render the whole video, so they take longer than trim and add audio.