Overview
VEED Fabric 1.0 is an image-to-video model by VEED for generating realistic talking avatars. By combining a static image and an audio track, it produces precise lip-sync videos where facial expressions and head movements follow the speech rhythm. It is well-suited for personalized marketing content, educational videos, and social ads. Creators often pair it with image generators like Nano Banana Pro or Seedream 4.0 to design custom characters before animating them.
Specifications · Text + image + audio to video
- Input mode
- Text + image + audio to video
- Accepts
- start frame, audio (required)
- Aspect ratios
- 16:9, 9:16, 4:3, 3:4, 1:1
- Resolutions
- 480p, 720p
- Native audio
- Audio-driven
Pricing
Build with this model: the VEED Fabric 1.0 API on the Hedra Developer Platform.
VEED Fabric 1.0 API →Real output · generated with VEED Fabric 1.0
Best of VEED Fabric 1.0
Prompting
Prompt tips
- Optimize the Source Image: Provide a high-resolution, forward-facing portrait with a neutral expression. Avoid images where hands or objects obscure the mouth and jawline.
- Clean Audio is Crucial: The model relies on clear phoneme detection. Ensure your audio track is free of background noise or heavy echo to prevent mouth jitter.
- Leverage Emotion Tags: If using text-to-speech, insert bracketed tags like
[excited],[whisper], or[confident]in your script to drive dynamic, sentence-level facial expressions. - Build Custom Spokespeople: Generate a unique character using an image model like Flux 1.1 Pro or Nano Banana, then use Fabric 1.0 to bring them to life with a cloned voice.
About the model
Questions, answered
What is VEED Fabric 1.0 best used for?
VEED Fabric 1.0 is an image-to-video model specialized in creating lip-synced talking avatars. It excels at turning a single static image—whether a photorealistic human, 3D mascot, clay figure, or illustration—and an audio file into a dynamic video. The model synchronizes mouth movements, head gestures, and body language to match the speech rhythm. It is widely used for marketing videos, educational content, and social media ads where you need realistic speech animation without filming.
Who created Fabric 1.0 and are there other versions?
Fabric 1.0 was developed by the video editing platform VEED and launched via API in mid-2025. Powered by a Diffusion Transformer (DiT) architecture, it processes visual and audio data simultaneously. VEED also offers a speed-optimized variant called VEED Fabric 1.0 Fast, which trades a small amount of visual fidelity for significantly faster generation times and lower inference costs.
How can I get the most realistic or expressive results from this model?
For the best lip-sync quality, ensure your input audio is clear and free of background noise. A popular community workflow involves generating a custom base character using image models like Nano Banana 2 or Nano Banana Pro before animating it with Fabric. Additionally, if you are using VEED's text-to-speech features, you can insert bracketed audio tags—such as [excited], [whisper], or [sigh]—directly into your script to control the avatar's emotional delivery and pacing at the sentence level.
