/\ ATFGraph /docs
README Downloads
Get started / ATFGraph

Framework reference

Find the right guide by what you want to do. Start with your first node for the managed installation → sign-in → build → deploy → Console workflow.

  • Examples Complete original examples, arranged from a first node to multi-agent systems. Each page includes its original prerequisites and execution details.
  • Framework concepts Understand the execution model before choosing capabilities or integrations.
  • Python SDK Build nodes with the ATFGraph Python SDK: skills, traversals, decisions, state and multi-node composition. Start with the hosted workflow in Develop & Deploy and the node manifest & configuration reference, then use these guides to add capabilities.
  • Guides and integrations Connect the framework to discovery protocols, tracing and authoring tools.
  • HTTP API & schemas HTTP operations and schemas for integrations. Ask AI-LAB to enable the access and capabilities your integration needs. For Python interfaces, use Python SDK; for hosted account and deployment commands, use CLI reference.

Choose the right kind of page

Python SDK explains how to build nodes. Examples show complete implementations. HTTP API & schemas documents integration contracts. The CLI reference covers installed hosted commands. For Kubernetes, self-hosting and additional integrations, see Platform capabilities and contact AI-LAB.

ATFGraph documentationHosted workflow · Framework reference