What Nomic AI does

Nomic AI is an open-source AI infrastructure company building tools that make embeddings, retrieval, and unstructured data exploration accessible to every developer. The company's flagship products are Nomic Atlas, a platform that lets teams visualize, search, and curate millions of documents and embeddings on interactive maps, and the Nomic Embed family of open-weights text and multimodal embedding models that consistently rank near the top of MTEB and Hugging Face leaderboards. Nomic also stewards GPT4All, the popular local LLM runtime that has been downloaded more than 250,000 times each month.

Nomic Embed models — including Nomic Embed Text v1.5, Nomic Embed Vision, and the multimodal Nomic Embed series — are released with fully open weights, training data, and training code, a deliberate choice that contrasts with the closed embedding APIs from OpenAI, Cohere, and Voyage. Atlas combines those embeddings with a managed vector index and a browser-based map UI so engineers can debug RAG retrieval, dedupe datasets, and spot drift in production.

Who it's for

Nomic AI targets AI engineers, data scientists, and ML platform teams at startups, research labs, and Fortune 500 enterprises who want best-in-class open embeddings without vendor lock-in. The product is especially popular with teams building RAG, search, and recommendation systems that need to inspect and curate massive embedding spaces.

Pricing

Nomic Atlas offers a free tier with usage caps, a Pro tier with higher quotas and team features, and an enterprise plan with self-hosted deployments, SSO, and dedicated support. Nomic Embed models are free to download and run under permissive open licenses, with paid hosted inference available through Atlas.

Team & funding

Nomic AI was founded in 2022 by Brandon Duderstadt (CEO) and Andriy Mulyar (CTO), both Johns Hopkins University engineers. The company raised a $2M pre-seed in 2022 led by Factorial Capital, followed by a $17M Series A in October 2024 led by Coatue at a reported $100M valuation, with participation from Betaworks, Contrary, MongoDB Ventures, and Factorial. Total funding stands at roughly $19M.

Position vs competitors

Nomic competes with closed embedding providers like OpenAI, Cohere, and Voyage AI (acquired by MongoDB) and with data exploration tools like Pinecone's data plane and Weaviate. Its differentiation is a fully open-source stack — weights, training data, and on-prem deployment — paired with the visual Atlas product that no closed competitor offers.