Crevid

Overview

Crevid is a web-based AI generation platform that combines image and video creation in a single interface. It accepts text prompts or image uploads as input and routes them through dedicated image or video models, depending on the user's selection.

According to the official documentation, Crevid offers a curated set of supported image models and supported video models, though the specific model names and architectures are not disclosed on the public homepage. This puts Crevid in the broader AI Photo & Image Generator category alongside tools like Flux Image and Supawork AI, but with the added dimension of video output that distinguishes it from image-only alternatives.

How It Works

The platform follows a four-step generation workflow. Users begin by typing a prompt or uploading an image. The vendor claims that a "Generate With AI" feature can expand a few keywords into a complete prompt, which the manufacturer describes as a way to achieve faster results. When enabled, a seed Lock button auto-generates a unique seed value to anchor the output for reproducibility — a feature commonly found in professional generative tools like Stable Diffusion.

Once configured, users click Generate. The vendor states that images complete in seconds, while videos can take a few minutes. Finished outputs appear directly in the interface and are available for download.

What's Supported

• Image models: Available for still-image generation from text or image inputs.
• Video models: Available for motion-content generation with longer processing times.
• Seed control: Lock and auto-seed for consistent, repeatable outputs.
• AI prompt expansion: Keyword-to-prompt assistance for faster ideation.
• Image upload input: Use existing images as generation starting points.
• Direct download: Results are downloadable once processing completes.

Limitations to Consider

The homepage is the sole source of information at this stage, and all performance claims — including image generation speed and prompt expansion quality — are vendor statements without independent verification. The specific image and video models are named only as categories, leaving potential users without information about resolution limits, style capabilities, or model provenance. Video generation latency of "a few minutes" may disrupt rapid-iteration workflows, and no pricing or tier information is available from the source packet.