Six model providers, embeddings, and self-hosted models for data that may not leave. ATX Agent Hub does not tie you to one model vendor. Each model call names a connection, and the connection decides the provider: Azure OpenAI, Azure AI Foundry, Google Gemini through AI Studio, Google Vertex AI with a service account, OpenAI or Anthropic. For data that must stay inside your network, administrators register self-hosted open-weight models served by llama.cpp. Every call reports provider-measured token usage, which flows into the run's cost.
This site needs JavaScript for the full experience. Every page above is reachable without it.