SUMMARY/ABSTRACT/SHORT DESCRIPTION
In the fast-paced ecosystem of AI, applications are often tightly coupled to specific vendors, making it difficult to swap models, vector stores, or tool runtimes as newer/better alternatives emerge. Enter OGX, the Open GenAI stack that can do X ("anything"), hence "OGX," an open source production-ready agentic OpenResponses-compliant API server. OGX acts as an interoperability "mux-like" universal translation layer that natively speaks the OpenAI, Anthropic, and Google AI APIs, letting you route inference, MCP, vector storage, file processing, and other requests to supported providers of your choice without updating a single line of code when your backend tools change. In this session, you will learn what OGX is, how it works, and how to get started building more flexible agents and AI applications. By the way, it's also devtool-agnostic, supporting Claude Code, Codex CLI, and OpenCode.

LENGTH
This is a 30- to 45-minute technical session talk aimed at AI developers, practioners, and engineering leadership.

AUDIENCE/WHO SHOULD ATTEND and PREREQUISITES [Beginner/Intermediate]
Backend software engineers, agent or AI/ML application developers, platform engineers, engineering leadership and IT decision makers

TAGS
OGX, open source, OSS, LLama Stack, OpenAI, Docling, inference, Ollama, vLLM, Meta, Google, Red Hat, AWS, Gemini, FAISS, Milvus, Qdrant, PGVector, Weviate, Python, Node.js, JavaScript, Typescript, agents, agentic, developers, API, APIs, JSON, tool, tools, Tavily, MCP, MCP servers, SQLite-Vec, ChromaDB, GenAI, web, desktop, mobile, developer, chunking, RAG, retrieval-augmented generation, vector databases, vector store, vector storage, embedding, files, local, cloud, HuggingFace, local, local filesystem, cloud storage