LangGraph Updates & Release Notes

Changes since sdk==0.3.14

• release(checkpoint): 4.1.1 (#7890)
• release(sdk-py): 0.3.15 (#7891)
• fix(sdk-py): percent-encode caller-supplied identifiers in URL paths (#7893)
• release(langgraph): 1.2.1 (#7883)
• chore(deps): bump idna from 3.11 to 3.15 in /libs/sdk-py (#7863)
• chore(deps): bump urllib3 from 2.6.3 to 2.7.0 in /libs/sdk-py (#7764)
• chore(deps): bump langsmith from 0.7.31 to 0.8.0 in /libs/sdk-py (#7789)
• release: bump alpha packages to official versions (#7775)
• chore(langgraph): bump langchain-core to 1.4.0 (#7767)
• feat(sdk-py): support metadata filter for crons search/count (#7737)
• chore(deps): bump ty from 0.0.23 to 0.0.33 in /libs/sdk-py (#7666)

Original source

Changes since 1.2.0a7

release: bump alpha packages to official versions (#7775)

feat(langgraph): durable error-handler resume across host crashes (#7773)

feat(langgraph): add set_node_defaults() to StateGraph (#7747)

chore(deps): bump urllib3 from 2.6.3 to 2.7.0 in /libs/langgraph (#7766)

chore(deps): bump mistune from 3.2.0 to 3.2.1 in /libs/langgraph (#7733)

chore(langgraph): bump langchain-core to 1.4.0 (#7767)

feat(checkpoint): force delta channel snapshot after max supersteps since last snapshot (#7746)

test(langgraph): de-flake heartbeat progress test (#7735)

chore(langgraph): re-implement exit mode for delta channel (#7730)

chore(deps): bump ty from 0.0.23 to 0.0.33 in /libs/sdk-py (#7666)

docs(checkpoint): mark DeltaChannel and delta-history APIs as beta (#7732)

chore(deps): bump jupyter-server from 2.17.0 to 2.18.0 in /libs/langgraph (#7713)

feat(checkpoint-sqlite): override get_delta_channel_history with streaming walk (#7702)

chore: "chore: minor clean up around checkpoint and delta channel" (#7706)

chore: minor clean up around checkpoint and delta channel (#7705)

Original source

Changes since prebuilt==1.1.0a2

release: bump alpha packages to official versions (#7775)

chore(deps): bump urllib3 from 2.6.3 to 2.7.0 in /libs/prebuilt (#7763)

chore(langgraph): bump langchain-core to 1.4.0 (#7767)

chore(deps): bump ty from 0.0.23 to 0.0.33 in /libs/sdk-py (#7666)

feat(checkpoint-sqlite): override get_delta_channel_history with streaming walk (#7702)

chore(deps-dev): bump the minor-and-patch group across 1 directory with 3 updates (#7698)

release: alpha bump (a4) for langgraph, checkpoint, checkpoint-postgres (#7701)

release: alpha bump langgraph 1.2.0a6 (#7697)

Original source

Changes since checkpoint==4.1.0a4

release: bump alpha packages to official versions (#7775)

chore(deps): bump urllib3 from 2.6.3 to 2.7.0 in /libs/checkpoint (#7762)

chore(deps): bump langchain-core from 1.3.2 to 1.3.3 in /libs/checkpoint (#7752)

feat(checkpoint): force delta channel snapshot after max supersteps since last snapshot (#7746)

fix(checkpoint): specify allowed_objects in Reviver (#7743)

chore: remove keepset helper (#7745)

chore(langgraph): add guide/conformance for delta channel checkpointer (#7736)

docs(checkpoint): mark DeltaChannel and delta-history APIs as beta (#7732)

chore(deps): bump the minor-and-patch group across 1 directory with 3 updates (#7670)

chore: "chore: minor clean up around checkpoint and delta channel" (#7706)

chore: minor clean up around checkpoint and delta channel (#7705)

Original source

Changes since sdk==0.3.13

release(sdk-py): 0.3.14 (#7712)

feat(sdk-py): add return_minimal to threads update (#7704)

release: alpha bump (a4) for langgraph, checkpoint, checkpoint-postgres (#7701)

release: alpha bump langgraph 1.2.0a6 (#7697)

release: alpha bump prebuilt 1.1.0a2, langgraph 1.2.0a5 (#7682)

release: alpha bump prebuilt 1.1.0a1, langgraph 1.2.0a4 (#7679)

feat(langgraph): dispatch stream_events(version='v3') on Pregel (#7677)

release: alpha bump (a3) for langgraph, checkpoint, checkpoint-postgres (#7678)

release: alpha for timers (#7647)

chore: update x links to langchain_oss (#7645)

feat(langgraph): add streaming transformer infrastructure and tests (#7519)

chore(deps): bump the minor-and-patch group across 1 directory with 4 updates (ty held back) (#7635)

release(prebuilt): 1.0.12, langgraph 1.1.10 (#7623)

release(checkpoint): 4.0.3 (#7625)

release(prebuilt): 1.0.11 (#7610)

feat(prebuilt): allow ToolNode tools to return list[Command | ToolMessage] (#7596)

chore(langgraph): bump version 1.1.8 -> 1.1.9 (#7563)

release(langgraph): 1.1.8 (#7545)

release(prebuilt): 1.0.10 (#7541)

release(langgraph): 1.1.7 (#7540)

chore(deps): bump langsmith from 0.7.20 to 0.7.31 in /libs/sdk-py (#7528)

release(checkpoint): 4.0.2 (#7518)

chore(deps-dev): bump pytest from 9.0.2 to 9.0.3 in /libs/sdk-py (#7504)

release(langgraph): 1.1.7a2 (#7511)

chore: allow passing some metadata only for tracing purposes (#7383)

release(langgraph): 1.1.7a1 (#7476)

chore(deps): bump langchain-core from 1.2.22 to 1.2.28 in /libs/sdk-py (#7449)

Original source

Changes since checkpointsqlite==3.0.3

feat(checkpoint-sqlite): override get_delta_channel_history with streaming walk (#7702)
release: alpha bump (a4) for langgraph, checkpoint, checkpoint-postgres (#7701)
feat: public get_writes_history saver API + delta cadence rework (#7699)
chore(deps): bump the minor-and-patch group in /libs/checkpoint-sqlite with 2 updates (#7668)
release: alpha bump (a3) for langgraph, checkpoint, checkpoint-postgres (#7678)
release: alpha for timers (#7647)
chore: update x links to langchain_oss (#7645)
release(checkpoint): 4.0.3 (#7625)
chore(deps): bump langsmith from 0.6.4 to 0.7.31 in /libs/checkpoint-sqlite (#7524)
release(checkpoint): 4.0.2 (#7518)
docs: document LANGGRAPH_STRICT_MSGPACK for checkpoint security (#7517)
chore(deps): bump pytest from 9.0.2 to 9.0.3 in /libs/checkpoint-sqlite (#7502)
chore(deps): bump langchain-core from 1.2.22 to 1.2.28 in /libs/checkpoint-sqlite (#7451)
chore(deps): bump the minor-and-patch group in /libs/checkpoint-sqlite with 3 updates (#7374)
chore(deps): bump pygments from 2.19.2 to 2.20.0 in /libs/checkpoint-sqlite (#7350)
chore(deps): bump langchain-core from 1.2.11 to 1.2.22 in /libs/checkpoint-sqlite (#7318)
chore(deps): bump requests from 2.32.5 to 2.33.0 in /libs/checkpoint-sqlite (#7279)
chore(deps): bump the all-dependencies group in /libs/checkpoint-sqlite with 2 updates (#7248)
chore(deps): bump ruff from 0.15.5 to 0.15.6 in /libs/checkpoint-sqlite in the all-dependencies group (#7191)
chore(deps): bump orjson from 3.11.5 to 3.11.6 in /libs/checkpoint-sqlite (#7144)
chore(deps): bump the all-dependencies group across 1 directory with 2 updates (#7070)
release(checkpoint): 0.4.1 (#6966)
chore: add serde events (#6954)
chore: update defaults (#6953)
release: rc2 (#6949)
release: Candidate (#6947)
Merge commit from fork
chore(deps-dev): bump ruff from 0.15.1 to 0.15.2 in /libs/checkpoint-sqlite in the all-dependencies group (#6913)
chore: add make type target for type checking (#6748)
chore(deps): bump ruff from 0.14.13 to 0.15.1 in /libs/checkpoint-sqlite in the all-dependencies group (#6807)
chore(deps): bump langchain-core from 1.2.7 to 1.2.11 in /libs/checkpoint-sqlite (#6828)

Original source

langgraph v1.2 (alpha)

This release adds finer-grained control over node execution — timeouts, error recovery, and graceful shutdown — a new channel type that cuts checkpoint overhead for long-running threads, and a new content-block-centric streaming API (v3) with typed, per-channel projections.

DeltaChannel (beta)

A new channel type that stores only the incremental delta at each step rather than re-serializing the full accumulated value. Most useful for channels that grow large over time — for example, a message list in a long-running thread. Without it, the full message list is written into every checkpoint; with DeltaChannel, only the new messages from each step are stored.

Use snapshot_frequency=K to write a full snapshot every K steps and bound read latency:

from typing import Annotated, Sequence
from typing_extensions import TypedDict
from langgraph.channels import DeltaChannel

# batch reducer: receives the current value and all writes from the superstep at once
def list_reducer(messages: list[str], writes: Sequence[list[str]]) -> list[str]:
    return [*messages, *(item for write in writes for item in write)]

class State(TypedDict):
    messages: Annotated[list[str], DeltaChannel(list_reducer, snapshot_frequency=5)]

Read the DeltaChannel docs →

Per-node timeouts

Pass timeout= to add_node to cap how long a single attempt may run. You can set a hard wall-clock limit (run_timeout), an idle limit that resets on progress (idle_timeout), or both. When the limit fires, LangGraph raises NodeTimeoutError, clears any writes from that attempt, and hands off to the retry policy. Timeouts apply to async nodes only; sync nodes with timeout= are rejected at compile time.

from langgraph.types import RetryPolicy, TimeoutPolicy

# Hard wall-clock limit: abort after 2 minutes regardless of progress
builder.add_node(
    "call_model",
    call_model,
    timeout=TimeoutPolicy(run_timeout=120),
    retry_policy=RetryPolicy(max_attempts=3),
)

# Idle limit: abort if no output is yielded for 30 seconds —
# useful for streaming nodes where partial progress should keep the clock reset
builder.add_node(
    "stream_response",
    stream_response,
    timeout=TimeoutPolicy(idle_timeout=30),
)

# Both: cap total duration and require steady progress
builder.add_node(
    "fetch_and_parse",
    fetch_and_parse,
    timeout=TimeoutPolicy(run_timeout=120, idle_timeout=30),
)

# Heartbeat: yield sentinel values to reset the idle clock during long async work,
# without writing anything meaningful to state
async def slow_tool(state: State) -> AsyncIterator[dict]:
    async for chunk in call_external_api():
        process(chunk)
        yield {}  # heartbeat — resets idle_timeout, writes nothing

builder.add_node(
    "slow_tool",
    slow_tool,
    timeout=TimeoutPolicy(idle_timeout=10),
)

Read the timeouts docs →

Node-level error handlers

Pass error_handler= to add_node to run a recovery function after all retries are exhausted. The handler receives a typed NodeError with the failing node name and exception, and can return a Command to update state and route to a different node — useful for Saga/compensation patterns where you want to recover gracefully rather than abort the graph. Handler failures propagate up normally; interrupt() inside a node bypasses the handler as expected.

from langgraph.errors import NodeError
from langgraph.types import Command, RetryPolicy

def payment_error_handler(state: State, error: NodeError) -> Command:
    return Command(
        update={"status": f"compensated: {error.error}"},
        goto="finalize",
    )

builder.add_node(
    "charge_payment",
    charge_payment,
    retry_policy=RetryPolicy(max_attempts=3, retry_on=ConnectionError),
    error_handler=payment_error_handler,
)

Read the error handler docs →

Graceful shutdown

Stop an in-flight run cooperatively — after the current superstep completes — and save a resumable checkpoint. Unlike timeouts, which fire mid-node, graceful shutdown waits for running work to finish before halting.

import signal
from langgraph.runtime import RunControl
from langgraph.errors import GraphDrained

control = RunControl()
signal.signal(signal.SIGTERM, lambda *_: control.request_drain("sigterm"))

try:
    result = graph.invoke(inputs, config, control=control)
except GraphDrained as e:
    # Stopped cleanly after the current superstep. Resume later with the same config.
    log.info("graph drained: %s", e.reason)

# Resume on next startup:
result = graph.invoke(None, config)

Read the graceful shutdown docs →

All five features require langgraph>=1.2. Timeouts and error handlers are Python-only; retry policies continue to work in both Python and TypeScript.

New event streaming API (beta)

A new content-block-centric streaming protocol that replaces dict-shaped events with typed, per-channel streams. version="v3" returns a GraphRunStream (sync) or AsyncGraphRunStream (async) — a handle whose typed projections the caller drives by iterating, rather than an iterator of StreamEvent dicts that callers must filter and reassemble.

run = graph.stream_events(input, version="v3")
for state in run.values:  # one snapshot per superstep
    print(state)
print(run.output)  # final state
print(run.interrupted, run.interrupts)

run.output, run.interrupted, and run.interrupts are populated regardless of which transformers are registered, so the "run to completion and read the result" path stays one line.

Per-LLM streams with typed sub-projections

run.messages yields one ChatModelStream (or AsyncChatModelStream) per LLM call. Each exposes typed sub-projections — text tokens, reasoning, tool calls, usage — so consumers can pick the slice they care about without filtering chunks by hand:

async for chat in graph.astream_events(input, version="v3").messages:
    async for token in chat.text:
        print(token, end="", flush=True)
tool_calls = await chat.tool_calls.collect()
final = await chat.output  # finalized AIMessage

This is the recommended path for token-level streaming to a UI in 1.2. Content-block streaming requires v3 end-to-end, including any inner graphs.

Built-in projections

Projection
What it carries

run.values
full state snapshot per superstep (also drives run.output etc.)

run.messages
one ChatModelStream per LLM call

run.lifecycle
subgraph started / completed / failed payloads

run.subgraphs
navigation handles for direct-child subgraphs

Other channels are opt-in — register them via compile(transformers=[...]) or the call-site transformers= on stream_events:

Transformer
Adds

UpdatesTransformer
run.updates

CustomTransformer
run.custom

CheckpointsTransformer
run.checkpoints

TasksTransformer
run.tasks

DebugTransformer
run.debug

from langgraph.stream import UpdatesTransformer, CustomTransformer

graph = builder.compile(transformers=[UpdatesTransformer, CustomTransformer])
run = graph.stream_events(input, version="v3")
for update in run.updates:
    print(update)

Projections are single-consumer by default — iterating run.values twice raises. Use projection.tee(n) (or atee(n) async) for fan-out, or run.interleave("messages", "values") to consume multiple projections in arrival order.

Custom transformers

Compile with transformers=[...] to register custom projections that propagate to subgraph scopes automatically. Each transformer declares the stream modes it needs; the runtime builds the union and subscribes to exactly that — no hardcoded defaults, no overhead for modes nobody consumes.

from typing import Any
from langgraph.stream import StreamTransformer, StreamChannel

class TokenCounter(StreamTransformer):
    """Surface per-call token usage for chat models at the run scope."""
    required_stream_modes = ("messages",)

    def __init__(self, scope: tuple[str, ...] = ()) -> None:
        super().__init__(scope)
        self._scope_list = list(scope)
        self.usage: StreamChannel[dict[str, int]] = StreamChannel()

    def init(self) -> dict[str, Any]:
        return {"token_usage": self.usage}

    def process(self, event) -> bool:
        if event["method"] != "messages":
            return True
        if event["params"]["namespace"] != self._scope_list:
            return True
        payload, _meta = event["params"]["data"]
        if not isinstance(payload, dict) or payload.get("event") != "message-finish":
            return True
        if "usage" in payload:
            self.usage.push(payload["usage"])
        return True

graph = builder.compile(transformers=[TokenCounter])
run = graph.stream_events(input, version="v3")
run.output  # drive to completion
usages = list(run.extensions["token_usage"])
total_in = sum(u.get("input_tokens", 0) for u in usages)
total_out = sum(u.get("output_tokens", 0) for u in usages)

Migration notes

v3 is opt-in via version="v3"; version="v1" and version="v2" are unchanged.

The protocol is in beta and may change before stabilization.

Original source