CLIP, which stands for Contrastive Language-Image Pretraining, is a neural network model developed by OpenAI that learns to understand the relationship between text and images by training on a vast dataset of image-text pairs. Rather than learning to generate images, CLIP learns to assess how well a given image matches a given text description, making it a powerful tool for evaluating, guiding, and interpreting visual content.

CLIP works by encoding both images and text into a shared embedding space, where semantically related items are positioned close together regardless of whether they are visual or textual. This means CLIP can compare an image of a sunset to the phrase "golden hour by the ocean" and assign a meaningful similarity score. This capability made CLIP foundational to early text-guided image generation systems, where it was used to steer the generative process toward outputs that matched a given prompt. Many influential image generation architectures from the early 2020s used CLIP guidance as a core component, and its influence persists across the broader landscape of multimodal AI.

For creators and practitioners working in AI generation, CLIP is relevant as background knowledge for understanding how models interpret and score prompts against visual output. Its role in text-image alignment underpins much of how modern AI generation systems respond to language, making it one of the foundational building blocks of the field.