# The infrastructure behind AI search in Figma

DevFeed: [The infrastructure behind AI search in Figma](<https://devfeed.tech/articles/the-infrastructure-behind-ai-search-in-figma-10123.md>)

Original publisher: [Read original article](<https://www.figma.com/blog/the-infrastructure-behind-ai-search-in-figma/>)

Author: Isaac Goldberg; Maggie Zhou

Published: 2024-10-15T00:00:00Z

Content type: article

Language: en

Sources: [Figma Blog](<https://devfeed.tech/sources/figma-blog.md>)

Topics: [AI search](<https://devfeed.tech/topics/ai-search.md>), [Figma](<https://devfeed.tech/topics/figma.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [User Interfaces](<https://devfeed.tech/topics/user-interfaces.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-search](<https://devfeed.tech/tags/ai-search.md>), [data](<https://devfeed.tech/tags/data.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [figma](<https://devfeed.tech/tags/figma.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [image](<https://devfeed.tech/tags/image.md>), [images](<https://devfeed.tech/tags/images.md>), [search](<https://devfeed.tech/tags/search.md>), [user-interfaces](<https://devfeed.tech/tags/user-interfaces.md>)

## AI overview

Figma describes the infrastructure behind its AI-powered search features for designs and published components. The system uses embeddings to support text-based and visual search across large collections of files and components, while addressing the cost of generating and indexing billions of embeddings. Figma says its embedding models were not trained on private Figma files or customer data; fine-tuning used images of user interfaces from public Community files.

## Source excerpt

To build AI search in Figma, we had to tackle a number of technical hurdles, including the challenge of generating and indexing billions of embeddings to power these features--while also keeping costs down.