Xpell is a layered runtime platform for AI-native systems. This page is a high-level map of the core parts, not deep package documentation.
Xpell is the AI-native platform behind Aime Technologies systems. It combines real-time execution, semantic memory, and a dynamic UI runtime so agents and applications can stay consistent while evolving.
These are the core building blocks that make Xpell a full platform rather than a thin model wrapper.
The runtime engine coordinates live agent behavior across reasoning steps, model actions, workflows, triggers, scheduling, and shared state transitions.
It is tuned for low-latency decisions that remain reliable across both client and server runtimes.
Capabilities
XDB stores entities, documents, embeddings, metadata, and relationships in one semantic data layer.
This foundation supports retrieval-augmented generation, contextual responses, personalization, and durable memory across sessions.
Capabilities
Xpell-UI (XUI) allows interfaces to be generated and updated through intent, schema, and runtime data instead of manual component wiring.
Xpell also includes a spatial runtime layer for 3D interfaces, built on three.js.
Capabilities
xnode runs Xpell on the backend and coordinates agents, data access, and real-time operations across services.
It provides the server-side runtime base required for full-stack AI-native applications.
Capabilities
Wormholes is Xpell's lightweight bidirectional communication layer that keeps runtime state synchronized in real time.
It propagates UI updates, agent actions, data changes, and workflow events between server and client.
Capabilities
Technical documentation, examples, and downloads live on xpell.ai.