Chroma is an open-source database for AI apps that stores documents with their embeddings and metadata, and finds the ones most similar to a query.
In plain words
Many AI apps need to fetch the most relevant pieces of text before asking a model a question. Chroma stores those pieces, their embeddings and extra labels called metadata. If you add plain documents, it can handle tokenizing, embedding and indexing for you, or you can supply your own embeddings. Searches can be dense, sparse or hybrid and can filter by metadata. It can run in memory for quick prototypes, as a server, or on the hosted Chroma Cloud.
A quick example
Chroma's own starter example adds two short documents, each tagged with where it came from, such as Notion or Google Docs. It then queries with a sentence and asks for the two closest results, optionally filtering by that source tag.
Where you'll meet it
- Retrieval for RAG chatbots.
- Prototyping semantic search in Python or JavaScript.
Related
Sources
- Docschroma-core/chroma, Chroma (GitHub) · read 29 Sept 2026
- DocsIntroduction, Chroma Docs · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.