Chroma Reaches Major Milestone: More Than One Million Computers Ran Its Vector Database for Building LLM AI Applications Last Month

SAN FRANCISCO, Nov. 01, 2023 (GLOBE NEWSWIRE) -- Chroma, the company building memory for AI, today announced that more than one million computers ran its open source database in the month of October. First released in February this year, Chroma now has more than 3.5 million downloads and 9,000 GitHub stars and is part of more than 10,000 public projects on GitHub. Chroma’s ease of use, large and active open source developer community, extensive integrations and relentless focus on serving AI applications has made it the favorite of developers worldwide.

"Vector databases are fast emerging as a critical link in the data pipelines that support generative AI initiatives. As companies build GenAI into their applications and workflows, they need to embed their own proprietary documents and text files into vector databases such as Chroma,” said Kevin Petrie, vice president of research at Eckerson Group. “Chroma’s open source platform, purpose built for AI, enables companies to feed the right content to language models. This improves the accuracy and value of GenAI outputs."

"We are honored that so many developers love the project that we, the community, have built together,” said Jeff Huber, co-founder and CEO of Chroma. “Together we are going to build some amazing things over the coming months.”

Chroma’s mission is to build the whole stack necessary to make reliable, robust, scalable memory a reality for every developer. 

The emergence of powerful general purpose LLMs into the mainstream has given rise to a new software stack which has made it possible to build practical applications with AI for the first time. In the new stack, the application logic layer uses LLMs programmed in natural language with total flexibility to perform tasks and answer queries. In order to serve the AI logic layer, the data layer, responsible for storage, memory and state requires a totally new approach. 

The first job of the data layer is to retrieve information relevant to the user’s query or task, with the same flexibility as the LLM itself. This means storing and retrieving data based on its semantic meaning, not just its content - this is accomplished by a vector database which is what Chroma’s open source product does today.

But beyond storing and retrieving information, the data layer must also provide developers with tools like embedding model selection, data chunking and signals about the relevancy of the available data. It must be a complete retrieval system, built from the ground up for AI. 

Chroma’s cloud platform will let developers access Chroma’s capabilities in a fully serverless, elastically scaling way, over the same simple API developers know and love.  

About Chroma
Chroma’s mission is to build the whole stack necessary to make reliable, robust, scalable memory a reality for every AI application developer. Chroma’s AI cloud-native open source vector database has been downloaded millions of times by developers building LLM-enabled applications. Chroma is backed by Quiet Capital and headquartered in San Francisco. For more on, visit or follow

Media and Analyst Contact:
Amber Rowland