Empowering Developers with
Open Source AI

Explore Vyva.ai's cutting-edge foundation models, inference libraries, and frameworks built for the community.

Foundation Models

State-of-the-art open weights released under permissible licenses.

LLM

Vyva-7B Instruct

A highly efficient 7 billion parameter model fine-tuned for complex instruction following, reasoning tasks, and enterprise applications.

7B Params Apache 2.0
Vision

Vyva-Vision Omni

Our multimodal model capable of understanding complex images, documents, and visual layouts with high accuracy.

Multimodal MIT License

Developer Libraries

Tools and frameworks to accelerate your AI adoption.

VyCache

A blazing fast distributed semantic caching layer for LLM applications. Reduces API costs and latency dramatically.

AgentFlow

An elegant framework for building complex, multi-agent workflows with built-in memory management and tool orchestration.