langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

[PR #2876] Add Node Caching Support #2970

New Issue
Closed
opened 2026-02-20 17:47:52 -05:00 by yindo · 0 comments
yindo commented 2026-02-20 17:47:52 -05:00
Owner
Copy Link
Copy Source

Original Pull Request: https://github.com/langchain-ai/langgraph/pull/2876

State: closed
Merged: No

Users of the StateGraph API can optionally cache the outputs of their nodes by providing a cache_policy to StateGraph::add_node.

builder = StateGraph(State)
builder.add_node(
    node_fn, 
    cache_policy=CachePolicy(
        key=operator.itemgetter("x"), # Mandatory
        ttl=NODE_CACHE_TTL,           # Optional
    ),
)
graph = builder.compile(checkpointer=checkpointer)

Where key is a one argument function that takes the input state of the graph and returns a cache_key as a string.

Important:

  • Node caching requires a checkpoint.
  • Caches are thread scoped

🪚 Sharp Edges 🪚:

  • If two graphs use the same runnable name, the same thread id, and the same cache key, there will be cache contamination 😷.
  • In the same multi-step process, Send calls with identical input will be deduplicated. However, in the V2 API, this behavior is more complex—the node might still execute because it is scheduled in the same iteration it was created, rather than the next one.
**Original Pull Request:** https://github.com/langchain-ai/langgraph/pull/2876 **State:** closed **Merged:** No --- Users of the `StateGraph` API can optionally cache the outputs of their nodes by providing a `cache_policy` to `StateGraph::add_node`. ``` builder = StateGraph(State) builder.add_node( node_fn, cache_policy=CachePolicy( key=operator.itemgetter("x"), # Mandatory ttl=NODE_CACHE_TTL, # Optional ), ) graph = builder.compile(checkpointer=checkpointer) ``` Where `key` is a one argument function that takes the input state of the graph and returns a `cache_key` as a _string_. Important: - Node caching requires a `checkpoint`. - Caches are `thread` scoped 🪚 Sharp Edges 🪚: - If two graphs use the same runnable name, the same thread id, and the same cache key, there will be cache contamination 😷. - In the same multi-step process, Send calls with identical input will be deduplicated. However, in the V2 API, this behavior is more complex—the node might still execute because it is scheduled in the same iteration it was created, rather than the next one.
yindo added the pull-request label 2026-02-20 17:47:52 -05:00
yindo closed this issue 2026-02-20 17:47:52 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#2970
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?