Agent Cloud

Agent Cloud provides a platform for building AI agents that answer questions using connected business data. The workflow combines source selection, data preparation, embedding, vector storage, and an agent configured with a chosen language model. Users can upload documents or connect supported systems, and database connections can be narrowed to selected tables and columns. File preparation includes basic splitting and semantic chunking options. Synchronization can run manually or on a schedule so that the agent works with refreshed source material. The platform describes support for local models with an OpenAI-compatible endpoint, as well as cloud providers. Self-managed deployments can connect to a locally hosted model; privacy therefore depends on the deployment and provider choices, rather than following automatically from the product name. Agent Cloud also includes access controls for teams using AI applications. Its official site links to the community repository and identifies an AGPL license. Teams evaluating it should account for the infrastructure required to run the data pipeline and for model or hosting costs when choosing between self-managed and hosted arrangements.