An AI model that understands or produces more than one type of data, such as text, images and audio together.
A multimodal model is not limited to text. It can read a photo of an invoice and extract its data, describe a product image, understand a voice note or read a screenshot of a design. Some models understand several modalities; some also produce them, such as generating images or speech.
This opens practical app ideas: a customer photographs a spare part to find the matching product, a field rep snaps a receipt that gets logged automatically, or a WhatsApp voice note becomes an order. When building apps, you can attach a design image so the result follows it. Practical tip: test with real images from your customers' world, not just perfect ones: poor lighting, Arabic handwriting, tilted photos. Never rely on automatic extraction for sensitive fields like amounts without a confirmation step from the user.