Advanced · 8 articles · MODULE 6: AI AGENTS
Tutorial · Advanced · 19 min
A checkpointer writes the graph's state to storage after every super-step, keyed by a thread_id you choose, which is the foundation for memory, resumption, and time travel.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 20 min
interrupt() stops a graph in the middle of a node, hands a payload to a reviewer, and persists everything so the run can resume later with Command(resume=value).
By: Anacodic Team
Start Reading →Tutorial · Advanced · 20 min
Because a checkpointer saves a snapshot after every super-step, a thread is an addressable chain of past states you can replay from or fork into a new branch.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 19 min
stream_mode chooses what a running graph emits: updates gives one event per node completion, values gives the whole state per super-step, and custom carries progress from inside a node.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 19 min
Send lets one node dispatch N parallel copies of another node when N is only known at runtime, merging their results through a reducer in a single super-step.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 19 min
Command lets a node return its state change and its next destination in one object, replacing the pair of a node returning a dict and a separate router function reading it.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 21 min
compile() returns a Runnable, so a compiled graph can be added as a node in another graph, and everything else follows from whether the two state schemas share keys.
By: Anacodic Team
Start Reading →Tutorial · Advanced · 20 min
A checkpointer persists one thread's working state; a Store persists facts across all threads, in namespaces that scope memory per user rather than per conversation.
By: Anacodic Team
Start Reading →