Neum AI

Neum AI is an open-source framework for building and managing data infrastructure for Retrieval Augmented Generation (RAG) and semantic search. It comes with a library of connectors to quickly set up data pipelines that transform unstructured and structured data into vector embeddings that are ready to be used to generate search indexes.

Neum AI can scale pipelines to handle millions of vectors and has a managed platform that keeps all vectors up to date as the underlying data changes. That's useful for large-scale and real-time data.

Some of the features include:

• Open-source SDKs: Build data flows with RAG-first frameworks that focus on the most important data transformations.
• Built-in connectors: Supports many data sources, embedding models and vector databases, with the ability to add custom connectors.
• Test and deploy: Run data pipelines locally and deploy them directly to the Neum AI cloud.
• Production-ready cloud platform: Offers a distributed architecture for embedding generation and ingestion, real-time syncing, observability, smart retrieval, self-improvement, and governance.

Neum AI supports real-time data embedding and indexing for RAG pipelines, letting you easily integrate with services like Supabase. You can build pipelines that read data from Supabase Storage, embed it with OpenAI, and store it in a vector database.

Pricing tiers are available for different needs:

• Starter (Free): Includes open-source connectors and tools, deployable pipelines to cloud, and limited scale.
• Pro ($500/month): Includes unlimited scale on common cloud, pipeline scheduling, real-time syncing, and priority support.
• Enterprise (Quote-based): Includes dedicated support, custom connectors, and infrastructure customized to your needs.
• Basic ($180/year), Business ($280/year), and Enterprise ($480/year): Additional tiers with different features and customization options.

Customers can get started with Neum AI through its SDK and cloud products. The framework is most useful for those building large-scale RAG pipelines and who need mature tools to manage and optimize their data infrastructure.

Published on June 13, 2024