Agent Server changelog ↗

langchain guide intermediate agents workflows

Original Documentation

Documentation Index#
Fetch the complete documentation index at: https://docs.langchain.com/llms.txt Use this file to discover all available pages before exploring further.

Subscribe: Our changelog includes an RSS feed that can integrate with Slack, email, Discord bots like Readybot or RSS Feeds to Discord Bot, and other subscription tools.

Agent Server is an API platform for creating and managing agent-based applications. It provides built-in persistence, a task queue, and supports deploying, configuring, and running assistants (agentic workflows) at scale. This changelog documents all notable updates, features, and fixes to Agent Server releases.

v0.7.59#

Updated the Redis queue to use zset with threads, reducing CPU usage by 25% and improving performance by eliminating unnecessary locking and optimizing indexes.

v0.7.58#

Upgraded langgraph-checkpoint to 4.0.0 in storage_postgres/uv.lock to address CVE-2026-27794, with adjustments for dependency pinning issues.

v0.7.57#

Fixed a regression preventing all users from creating crons associated with system graphs.

v0.7.56#

Added support for ttl, index, and refresh_ttl parameters in store HTTP API endpoints to align with the SDK and in-process store interface.
Added support for the ?include=ttl query parameter in the GET /threads/{thread_id} endpoint to return TTL information.
Updated metrics reporting to accurately account for PostgreSQL and Redis connections, ensuring consistent statistics between GRPC and Python metrics.

v0.7.55#

Fixed a bug that caused duplicate run scheduling in the new cron scheduler backend.
Refactored the GET /docs endpoint to read from a static OpenAPI spec for improved compatibility with custom ingress configurations.

v0.7.54#

Resolved an issue where the custom encryption context for gRPC services was not loading correctly due to hardcoded values.

v0.7.52#

Added feedback URLs to the /wait and /join endpoint responses under a __feedback__ key when feedback_keys are supplied.
Added feedback_keys support to the distributed runtime, including presigned feedback token generation using langsmith-go SDK.
Upgraded Werkzeug from version 3.1.5 to 3.1.6 to address a Windows security issue with special device names in multi-segment paths.
Upgraded Go runtime to 1.25.7 to address critical and high severity CVEs identified in vulnerability scans.

v0.7.51#

Improved license check resilience during upstream outages, including a cached fallback, a 24-hour grace period, and automatic cleanup of Redis entries.
Ensured assistant descriptions and names are synced on startup when LANGSERVE_GRAPHS config changes.
Enhanced the Checkpointer API by introducing a two-level protocol hierarchy and fixing capability detection to support extended methods directly.

v0.7.49#

Reserved metadata keys in request payloads are now silently stripped rather than causing a 422 error, enhancing user experience.
Fixed an issue where store default TTL was not applied to items written without an explicit TTL parameter.

v0.7.46#

Structured error payloads in webhooks now include error and message fields, replacing the previous flat string format, which may affect systems parsing the error field.
Expanded store auth tests with namespace-rewriting to enhance namespace handling and cross-user isolation.

v0.7.45#

Replaced null bytes with U+FFFD in all handling paths to prevent key collisions.

v0.7.44#

Increase flexibility of database URI parser
Add additional validation on some payloads

v0.7.40#

Fixed a regression in assistant creation by ensuring metadata and config are populated as empty objects {} instead of null when not provided.

v0.7.39#

Ensured auth configuration is passed correctly in distributed runtime operations to improve executor functionality.
Added support for Red Hat UBI-9 based Docker images for enterprise customers using RHEL-based containers.
Added graceful shutdown handoff for distributed runtime, allowing in-flight runs to transfer to the next pod without using a retry attempt.

v0.7.38#

Added state_updated_at field to threads for tracking meaningful state changes, allowing filtering and sorting based on these changes.
Added support for scheduling crons within the core system.
Ensured accurate display of the https protocol in agent cards using the x-forwarded-proto header for proper A2A client functionality.

v0.7.37#

Added generic fallbacks for acopy_thread, aprune, and adelete_for_runs in the BYOC checkpointer adapter, simplifying implementation for custom checkpointers.

v0.7.36#

Updated A2A protocol support to v1.0 RC, renamed JSON-RPC methods, added a ListTasks handler, and enhanced role, state, and part formats for improved integration and compliance.
Improved authentication filtering for Crons.search() and Crons.count() to prevent unauthorized thread information access.
Fixed gaps in BYOC checkpointer for copy, rollback, and namespace filtering operations, ensuring proper handling across different storage backends.

v0.7.35#

Added an optional context parameter to MCP tools/call and A2A message/send and message/stream endpoints, enabling middleware to inject runtime context from headers.

v0.7.33#

Enhanced the Redis fixture by removing custom checkpointer test skips, improving typed serialization, and adding missing Redis methods.
Resolved a stored XSS vulnerability in the handle_ui endpoint by sanitizing message names in single-quoted HTML onload attributes.
Fixed an authorization bypass issue in put_item to ensure correct namespace rewrite by the auth handler.
Enforced assistant ownership checks during run creation, preventing execution on unowned assistants while ensuring system assistants remain accessible to all authenticated users.
Implemented AES encryption for checkpoint blobs and writes in the Go checkpointer when using LANGGRAPH_AES_KEY.
Implemented initial checkpointer gRPC servicer with all necessary methods and conversion helpers.

v0.7.32#

Sanitized error messages in streams and A2A responses to protect sensitive information like database connection strings and internal hostnames.
Fixed a bug in join_run_stream to correctly handle multiple stream_mode parameters, ensuring proper parsing of stringified JSON lists.
Added build, test, and publish processes for Node.js 24 images, supporting the latest LTS version.
Enhanced custom checkpointer adapter with new capabilities and improved metadata enrichment for consistent API responses.
Added stricter version constraints for langgraph libraries in executor Docker images to prevent unintended upgrades.
Enhanced security by sanitizing SSE event and id fields to prevent CR/LF injection.
Fixed an issue causing cron-created runs to use default encryption contexts instead of properly propagating the specified ones.

v0.7.31#

Corrected metadata reading functionality to ensure accurate data processing.

v0.7.30#

Propagated cron metadata for more comprehensive scheduling information.
Merges cron metadata on PATCH requests to align with other endpoints by preserving existing data.

v0.7.29#

Refined authentication semantics for cron creation to prevent privilege escalation and ensure independent filtering for crons, threads, and assistants.
Validated tar file entries to prevent directory traversal vulnerabilities in the cloudflared download process.
Added an IDs filter to the SearchThreadsRequest to streamline thread endpoint operations.
Updated the fallback mechanism to use a Python Postgres connection for thread state, fixing issues with worker completion checkpoints.
Introduced a feature-flagged initial version of the Redis queue implementation with ongoing updates.

v0.7.28#

Internal maintenance and stability improvements for MCP and gRPC.

v0.7.27#

Improved MCP tool input schemas by removing common message types for cleaner tool definitions.
Added name sanitization for MCP tools to ensure valid tool names.

v0.7.26#

Added validation for system keys on ingress.

v0.7.25#

Switched the Python queue worker to use core go Runs.next().
Fixed a monitoring issue in the long query monitor

v0.7.24#

Optimized Postgres connection handling to prevent hitting connection limits under high load and removed unnecessary error logs.
Switched to a new backend for runs management and streaming using gRPC.

v0.7.23#

Corrected the unmarshaling process for the input field in RunCommand to ensure accurate data mapping and enable a previously gated JS test.
Ensured race condition handling for run streaming by fully subscribing before execution starts, with added support for the FF_LOG_DROPPED_EVENTS environment variable.

v0.7.22#

Ensured get_store() works in custom routes, enabling Store access from user-defined Starlette endpoints.

v0.7.21#

Support for PATCH /crons/{cron_id}

v0.7.19#

Custom encryption improvements for core API.

v0.7.18#

Update thread streaming for core API usage.

v0.7.17#

Instrumentation for OTEL now requires explicit opt-in with LS_APM_OTEL_ENABLED=true for improved control.

v0.7.16#

Switched threads streaming to the new gRPC backend for improved performance.
Introduced replica tracing in DR for enhanced evaluation capabilities in Studio.

v0.7.15#

Added support for pausing crons with a new is_enabled field, allowing only enabled crons to be executed.
Introduced gRPC server support for JSON encryption and decryption operations.

v0.7.14#

Ensured selected system fields are excluded from custom encryption to prevent unnecessary encryption of non-sensitive data.
Introduced a custom checkpointer adapter with unit tests to validate implementation checks.

v0.7.13#

Fixed a bug where the app state was not properly preserved through requests when a mount prefix was set.

v0.7.11#

Added configuration to control which payload fields can be exposed to webhooks.
Updated all dependencies in the /api/langgraph_api/js group, including @langchain/core, hono, @types/react, and prettier, to the latest versions for improved performance and security.
Upgraded hono from version 4.11.4 to 4.11.7 to address multiple security vulnerabilities in the middleware.

v0.7.10#

Increased the gRPC server startup timeout to 1 minute to prevent occasional connection timeouts with the core server.
Updated @langchain/langgraph from version 1.1.0 to 1.1.2, introducing mixed schema support for StateGraph and type bag patterns for GraphNode and ConditionalEdgeRouter utilities.

v0.7.9#

A2A messageId is now mapped to LangChain message IDs for proper message tracking across protocols.

v0.7.7#

Ensured preservation of custom configurable fields in checkpoint metadata during gRPC serialization.

v0.7.5#

Enforced custom encryption for values, interrupts, and errors when setting a thread’s status, resolving previous inconsistencies.
Added A2A validation checks in message/stream and message/send routes for parts, role, and messageId fields.
Added native A2A interrupt support: input-required state is now returned when graphs are interrupted. Use the new command parameter in message/stream and message/send requests to resume with a Command payload.
Mounted .well-known/agent-card.json under /a2a/{assistant_id}/ for A2A agent discovery.
Added proper A2A error codes for task existence checks in tasks/cancel.

v0.7.4#

Fixed a bug with Redis URL parsing for ssl_cert_reqs field, ensuring compatibility with redis-go.
Added a gRPC client for streaming runs, controlled by the FF_USE_CORE_API feature flag.

v0.7.2#

Updated @langchain/langgraph to version 1.1.0, introducing type utilities for graph nodes and conditional edges for enhanced TypeScript ergonomics.

v0.7.0#

Switched to using the Go assistants implementation by default for improved performance.
Added LANGGRAPH_AES_JSON_KEYS configuration to enable AES encryption for specified JSON fields using a key name allowlist.

v0.6.39#

Added gRPC client support for Threads.State() to the Python core-api, improving thread ID and run counting operations.

v0.6.36#

Validated the length of $and and $or in auth filters and optimized unnecessary root-level filters.

v0.6.35#

Unified the error format by removing the code field and standardizing all errors to return JSON with a detail field.

v0.6.34#

Small fixes for feature-flagged internal environments (unreleased).

v0.6.33#

Small fixes for feature-flagged internal environments (unreleased).

v0.6.32#

Small fixes for feature-flagged internal environments (unreleased).

v0.6.31#

Properly respected the disable_a2a setting to ensure accurate configuration handling.

v0.6.29#

Fix minor bugs.

v0.6.28#

Added support for ParentCommand to correctly propagate control to parent graphs, enhancing command handling and navigation.
Added a Python gRPC client for managing run operations, enhancing consistency between Go and Python implementations.

v0.6.27#

Fixed a regression issue in handling empty thread metadata.

v0.6.26#

Fixed the port configuration issue for the persistence gRPC server.

v0.6.25#

Ran the core-api gRPC server in the executor tier to support loopback API calls in graphs and removed unnecessary configuration for disabling the server.

v0.6.24#

Fixed the behavior of the liveness probe in the executor tier, addressing issues from version 0.6.23.

v0.6.23#

Integrated gRPC server health check with /ok endpoint in liveness probe to ensure proper startup coordination.
Reverted the previous change to disable the checkpointer and added a condition to enable RemoteCheckpointer only during testing.
Suppressed langgraph_auth_* and langgraph_request_id fields in checkpoint metadata to prevent inclusion of transient user data.

v0.6.22#

Resolved an error caused by missing encryption contexts when using blob-only custom encryption, ensuring proper function without errors.

v0.6.21#

Introduced a Python gRPC client for run operations, including Search, Get, Delete, Cancel, Stats, and Sweep, with updated API implementation and a new unit test suite for enum mappings.

v0.6.19#

Reproduced OSS implementations of get_state and update_state in the engine server and re-enabled test_weather_subgraph.

v0.6.18#

Added functionality to enforce specific license claims for self-hosted Enterprise users, enabling remote disabling of the Agent Builder product.
Added a new Prune endpoint for better resource management.
Merged graph configuration with invoke configuration in Pregel, giving precedence to invoke settings.
Introduced the include=ttl query parameter to the GET /threads/{id} endpoint for optional TTL information retrieval without affecting standard read performance.
Introduced a keep_latest TTL strategy to preserve the latest state while pruning older checkpoints via the core API.

v0.6.17#

Ensured ongoing runs are stopped when an agent is deleted to prevent lingering processes.

v0.6.16#

Streamlined and consolidated run operations in the Go persistence layer, improving efficiency and consistency across packages.

v0.6.15#

Improved the utility converting custom route docstrings to OpenAPI schema content by adding error handling when parsing docstrings, applicable for users with custom Starlette apps.

v0.6.12#

Improved resolve_embeddings to be more robust, enabling multiple calls without errors.
Updated @langchain/langgraph from version 1.0.4 to 1.0.7, adding support for resumableStreams on remote graphs and undeprecating toolsCondition.
Implemented RemoteCheckpointer to enable subgraph checkpointing, enhancing task execution reliability.

v0.6.11#

Made the maximum number of retries configurable for enhanced customization.

v0.6.10#

Ensured run cancellation only processes ‘message’ type Redis events, improving pubsub client reliability.
Added custom encryption for the Store API value field, allowing users to choose which keys to encrypt for enhanced security.
Enabled streaming for subgraph custom events by updating TeeStream to handle event types separately.

v0.6.9#

Enforced stable JSON keys for custom encryption, removed model-type-specific custom JSON functions, and improved error handling for double-encryption scenarios.

v0.6.8#

Added profiling feature to enhance performance analysis and monitoring.

v0.6.7#

Logged server startup time for improved monitoring and diagnostics.

v0.6.5#

Added a warning log that triggers during import time for improved visibility.

v0.6.4#

Enhanced custom encryption by parallelizing metadata and config processes, added encryption for thread.config and some checkpoints, improved tests and schema consistency.
Ensured the Go server starts as core-api in the queue entrypoint for consistent runtime behavior.

v0.6.2#

Resolved an issue that caused duplicate calls to middleware when mount_prefix was specified.

v0.6.0#

This minor version updates the streaming APIs /join-stream and /stream behavior with respect to the last-event-id parameter to align with the SSE spec. Previously, passing a last-event-id would return that message in addition to any following messages. Going forward, these APIs will only return new messages following the provided last-event-id. For example, with the following stream, previously passing a last-event-id of 2 would return the messages with ids 2 and 3, but will now only return the message with id 3:

{
    "id": 1,
    "event": "message",
    "data": {
        "content": "Excluded"
    }
},
{
    "id": 2,
    "event": "message",
    "data": {
        "content": "Passed last-event-id"
    }
},
{
    "id": 3,
    "event": "message",
    "data": {
        "content": "Included"
    }
}

This bump also includes some fixes, including a bug exposing unintended internal events in run streams.

v0.5.42#

Modified the Go server to rely solely on the CLI -service flag for determining service mode, ignoring the globally set FF_USE_CORE_API for better deployment specificity.

v0.5.41#

Fixed an issue with cron jobs in hybrid mode by ensuring proper initialization of the ENTERPRISE_SAAS global flag.

v0.5.39#

Completed the implementation of custom encryptions for runs and crons, along with simplifying encryption processes.
Introduced support for streaming subgraph events in both values and updates stream modes.

v0.5.38#

Implemented complete custom encryption for threads, ensuring all thread data is properly secured and encrypted.
Ensured Redis attempt flags are consistently expired to prevent stale data.
Added core authentication and support for OR/AND filters, enhancing security and flexibility.

v0.5.37#

Added a name parameter to the assistants count API for improved search flexibility.

v0.5.36#

Introduced configurable webhook support, allowing users to customize submitted webhooks and headers.
Added an /ok endpoint at the root for easier health checks and simplified configuration.

v0.5.34#

Introduced custom encryption middleware, allowing users to define their own encryption methods for enhanced data protection.

v0.5.33#

Set Uvicorn’s keep-alive timeout to 75 seconds to prevent occasional 502 errors and improve connection handling.

v0.5.32#

Introduced OpenTelemetry telemetry agent with support for New Relic integration.

v0.5.31#

Added Py-Spy profiling for improved analysis of deployment performance, with some limitations on coverage.

v0.5.30#

Always configure loopback transport clients to enhance reliability.
Ensured authentication headers are passed for remote non-stream methods in JS.

v0.5.28#

Introduced a faster, Rust-based implementation of uuid7 to improve performance, now used in langsmith and langchain-core.
Added support for $or and $and in PostgreSQL auth filters to enable complex logic in authentication checks.
Capped psycopg and psycopg-pool versions to prevent infinite waiting on startup.

v0.5.27#

Ensured runs.list with filters returns only run fields, preventing incorrect status data from being included.
(JS) Updated uuid from version 10.0.0 to 13.0.0. and exit-hook from version 4.0.0 to 5.0.1.

v0.5.26#

Resolved issues with store.put when used without AsyncBatchedStore in the JavaScript environment.

v0.5.25#

Introduced the ability to search assistants by their name using a new endpoint.
Casted store_get return types to tuple in JavaScript to ensure type consistency.

v0.5.24#

Added executor metrics for Datadog and enhanced core stream API metrics for better performance tracking.
Disabled Redis Go maintenance notifications to prevent startup errors with unsupported commands in Redis versions below 8.

v0.5.20#

Resolved an error in the executor service that occurred when handling large messages.

v0.5.19#

Upgraded built-in langchain-core to version 1.0.7 to address a prompt formatting vulnerability.

v0.5.18#

Introduced persistent cron threads with on_run_completed: {keep,delete} for enhanced cron management and retrieval options.

v0.5.17#

Enhanced task handling to support multiple interrupts, aligning with open-source functionality.

v0.5.15#

Added custom JSON unmarshalling for Resume and Goto commands to fix map-style null resume interpretation issues.

v0.5.14#

Ensured pg make start command functions correctly with core-api enabled.

v0.5.13#

Support include and exclude (plural form key for includes and excludes) since a doc incorrectly claimed support for that. Now the server accepts either.

v0.5.11#

Ensured auth handlers are applied consistently when streaming threads, aligning with recent security practices.
Bumped undici dependency from version 6.21.3 to 7.16.0, introducing various performance improvements and bug fixes.
Updated p-queue from version 8.0.1 to 9.0.0, introducing new features and breaking changes, including the removal of the throwOnTimeout option.

v0.5.10#

Implemented healthcheck calls in the queue /ok handler to improve Kubernetes liveness and readiness probe compatibility.

v0.5.9#

Resolved an issue causing an “unbound local error” for the elapsed variable during a SIGINT interruption.
Mapped the “interrupted” status to A2A’s “input-required” status for better task status alignment.

v0.5.8#

Ensured environment variables are passed as a dictionary when starting langgraph-ui for compatibility with uvloop.
Implemented CRUD operations for runs in Go, simplifying JSON merges and improving transaction readability, with PostgreSQL as a reference.

v0.5.7#

Replaced no-retry Redis client with a retry client to handle connection errors more effectively and reduced corresponding logging severity.

v0.5.6#

Added pending time metrics to provide better insights into task waiting times.
Replaced pb.Value with ChannelValue to streamline code structure.

v0.5.5#

Made the Redis health_check_interval more frequent and configurable for better handling of idle connections.

v0.5.4#

Implemented ormsgpack with OPT_REPLACE_SURROGATES and updated for compatibility with the latest FastAPI release affecting custom authentication dependencies.

v0.5.2#

Added retry logic for PostgreSQL connections during startup to enhance deployment reliability and improved error logging for easier debugging.

v0.5.1#

Resolved an issue where persistence was not functioning correctly with LangChain.js’s createAgent feature.
Optimized assistants CRUD performance by improving database connection pooling and gRPC client reuse, reducing latency for large payloads.

v0.5.0#

This minor version now requires langgraph-checkpoint versions later than 3.0 to prevent a deserialization vulnerability in earlier versions of the langgraph-checkpoint library. The langgraph-checkpoint library is compatible with langgraph minor versions 0.4, 0.5, 0.6, and 1.0.

This version removes default support for deserialization of payloads saved using the “json” type, which has never been the default. By default, objects are serialized using msgpack. Under certain uncommon situations, payloads were serialized using an older “json” mode. If those payloads contained custom python objects, those will no longer be deserializable unless you provide a serde config:

{
    "checkpointer": {
        "serde": {
            "allowed_json_modules": [
                ["my_agent", "my_file", "SomeType"],
            ]
        }
    }
}

v0.4.47#

Validated and auto-corrected environment configuration types using TypeAdapter.
Added support for LangChain.js and LangGraph.js version 1.x, ensuring compatibility.
Updated hono library from version 4.9.7 to 4.10.3, addressing a CORS middleware security issue and enhancing JWT audience validation.
Introduced a modular benchmark framework, adding support for assistants and streams, with improvements to the existing ramp benchmark methodology.
Introduced a gRPC API for core threads CRUD operations, with updated Python and TypeScript clients.
Updated hono package from version 4.9.7 to 4.10.2, including security improvements for JWT audience validation.
Updated hono dependency from version 4.9.7 to 4.10.3 to fix a security issue and improve CORS middleware handling.
Introduced basic CRUD operations for threads, including create, get, patch, delete, search, count, and copy, with support for Go, gRPC server, and Python and TypeScript clients.

v0.4.46#

Added an option to enable message streaming from subgraph events, giving users more control over event notifications.

v0.4.45#

Implemented support for authorization on custom routes, controlled by the enable_custom_route_auth flag.
Set default tracing to off for improved performance and simplified debugging.

v0.4.44#

Used Redis key prefix for license-related keys to prevent conflicts with existing setups.

v0.4.43#

Implemented a health check for Redis connections to prevent them from idling out.

v0.4.40#

Prevented duplicate messages in resumable run and thread streams by addressing a race condition and adding tests to ensure consistent behavior.
Ensured that runs don’t start until the pubsub subscription is confirmed to prevent message drops on startup.
Renamed platform from langgraph to improve clarity and branding.
Reset PostgreSQL connections after use to prevent lock holding and improved error reporting for transaction issues.

v0.4.39#

Upgraded hono from version 4.7.6 to 4.9.7, addressing a security issue related to the bodyLimit middleware.
Allowed customization of the base authentication URL to enhance flexibility.
Pinned the ’ty’ dependency to a stable version using ‘uv’ to prevent unexpected linting failures.

v0.4.38#

Replaced LANGSMITH_API_KEY with LANGSMITH_CONTROL_PLANE_API_KEY to support hybrid deployments requiring license verification.
Introduced self-hosted log ingestion support, configurable via SELF_HOSTED_LOGS_ENABLED and SELF_HOSTED_LOGS_ENDPOINT environment variables.

v0.4.37#

Required create permissions for copying threads to ensure proper authorization.

v0.4.36#

Improved error handling and added a delay to the sweep loop for smoother operation during Redis downtime or cancellation errors.
Updated the queue entrypoint to start the core-api gRPC server when FF_USE_CORE_API is enabled.
Introduced checks for invalid configurations in assistant endpoints to ensure consistency with other endpoints.

v0.4.35#

Resolved a timezone issue in the core API, ensuring accurate time data retrieval.
Introduced a new middleware_order setting to apply authentication middleware before custom middleware, allowing finer control over protected route configurations.
Logged the Redis URL when errors occur during Redis client creation.
Improved Go engine/runtime context propagation to ensure consistent execution flow.
Removed the unnecessary assistants.put call from the executor entrypoint to streamline the process.

v0.4.34#

Blocked unauthorized users from updating thread TTL settings to enhance security.

v0.4.33#

Improved error handling for Redis locks by logging LockNotOwnedError and extending initial pool migration lock timeout to 60 seconds.
Updated the BaseMessage schema to align with the latest langchain-core version and synchronized build dependencies for consistent local development.

v0.4.32#

Added a GO persistence layer to the API image, enabling GRPC server operation with PostgreSQL support and enhancing configurability.
Set the status to error when a timeout occurs to improve error handling.

v0.4.30#

Added support for context when using stream_mode="events" and included new tests for this functionality.
Added support for overriding the server port using $LANGGRAPH_SERVER_PORT and removed an unnecessary Dockerfile ARG for cleaner configuration.
Applied authorization filters to all table references in thread delete CTE to enhance security.
Introduced self-hosted metrics ingestion capability, allowing metrics to be sent to an OTLP collector every minute when the corresponding environment variables are set.
Ensured that the set_latest function properly updates the name and description of the version.

v0.4.29#

Ensured proper cleanup of redis pubsub connections in all scenarios.

v0.4.28#

Added a format parameter to the queue metrics server for enhanced customization.
Corrected MOUNT_PREFIX environment variable usage in CLI for consistency with documentation and to prevent confusion.
Added a feature to log warnings when messages are dropped due to no subscribers, controllable via a feature flag.
Added support for Bookworm and Bullseye distributions in Node images.
Consolidated executor definitions by moving them from the langgraph-go repository, improving manageability and updating the checkpointer setup method for server migrations.
Ensured correct response headers are sent for a2a, improving compatibility and communication.
Consolidated PostgreSQL checkpoint implementation, added CI testing for the /core directory, fixed RemoteStore test errors, and enhanced the Store implementation with transactions.
Added PostgreSQL migrations to the queue server to prevent errors from graphs being added before migrations are performed.

v0.4.27#

Replaced coredis with redis-py to improve connection handling and reliability under high traffic loads.

v0.4.24#

Added functionality to return full message history for A2A calls in accordance with the A2A spec.
Added a LANGGRAPH_SERVER_HOST environment variable to Dockerfiles to support custom host settings for dual stack mode.

v0.4.23#

Use a faster message codec for redis streaming.

v0.4.22#

Ported long-stream handling to the run stream, join, and cancel endpoints for improved stream management.

v0.4.21#

Added A2A streaming functionality and enhanced testing with the A2A SDK.
Added Prometheus metrics to track language usage in graphs, middleware, and authentication for improved insights.
Fixed bugs in Open Source Software related to message conversion for chunks.
Removed await from pubsub subscribes to reduce flakiness in cluster tests and added retries in the shutdown suite to enhance API stability.

v0.4.20#

Optimized Pubsub initialization to prevent overhead and address subscription timing issues, ensuring smoother run execution.

v0.4.19#

Removed warnings from psycopg by addressing function checks introduced in version 3.2.10.

v0.4.17#

Filtered out logs with mount prefix to reduce noise in logging output.

v0.4.16#

Added support for implicit thread creation in a2a to streamline operations.
Improved error serialization and emission in distributed runtime streams, enabling more comprehensive testing.

v0.4.13#

Monitored queue status in the health endpoint to ensure correct behavior when PostgreSQL fails to initialize.
Addressed an issue with unequal swept ID lengths to improve log clarity.
Enhanced streaming outputs by avoiding re-serialization of DR payloads, using msgpack byte inspection for json-like parsing.

v0.4.12#

Ensured metrics are returned even when experiencing database connection issues.
Optimized update streams to prevent unnecessary data transmission.
Upgraded hono from version 4.9.2 to 4.9.6 in the storage_postgres/langgraph-api-server for improved URL path parsing security.
Added retries and an in-memory cache for LangSmith access calls to improve resilience against single failures.

v0.4.11#

Added support for TTL (time-to-live) in thread updates.

v0.4.10#

In distributed runtime, update serde logic for final checkpoint -> thread setting.

v0.4.9#

Added support for filtering search results by IDs in the search endpoint for more precise queries.
Included configurable headers for assistant endpoints to enhance request customization.
Implemented a simple A2A endpoint with support for agent card retrieval, task creation, and task management.

v0.4.7#

Stopped the inclusion of x-api-key to enhance security.

v0.4.6#

Fixed a race condition when joining streams, preventing duplicate start events.

v0.4.5#

Ensured the checkpointer starts and stops correctly before and after the queue to improve shutdown and startup efficiency.
Resolved an issue where workers were being prematurely cancelled when the queue was cancelled.
Prevented queue termination by adding a fallback for cases when Redis fails to wake a worker.

v0.4.4#

Set the custom auth thread_id to None for stateless runs to prevent conflicts.
Improved Redis signaling in the Go runtime by adding a wakeup worker and Redis lock implementation, and updated sweep logic.

v0.4.3#

Added stream mode to thread stream for improved data processing.
Added a durability parameter to runs for improved data persistence.

v0.4.2#

Ensured pubsub is initialized before creating a run to prevent errors from missing messages.

v0.4.0#

Minor version 0.4 comes with a number of improvements as well as some breaking changes.

Emitted attempt messages correctly within the thread stream.
Reduced cluster conflicts by using only the thread ID for hashing in cluster mapping, prioritizing efficiency with stream_thread_cache.
Introduced a stream endpoint for threads to track all outputs across sequentially executed runs.
Made the filter query builder in PostgreSQL more robust against malformed expressions and improved validation to prevent potential security risks.

This minor version also includes a couple of breaking changes to improve the usability and security of the service:

In this minor version, we stop the practice of automatically including headers as configurable values in your runs. You can opt-in to specific patterns by setting configurable_headers in your agent server config.
Run stream event IDs (for resumable streams) are now in the format of ms-seq instead of the previous format. We retain backwards compatibility for the old format, but we recommend using the new format for new code.

v0.3.4#

Added custom Prometheus metrics for Redis/PG connection pools and switched the queue server to Uvicorn/Starlette for improved monitoring.
Restored Wolfi image build by correcting shell command formatting and added a Makefile target for testing with nginx.

v0.3.3#

Added timeouts to specific Redis calls to prevent workers from being left active.
Updated the Golang runtime and added pytest skips for unsupported functionalities, including initial support for passing store to node and message streaming.
Introduced a reverse proxy setup for serving combined Python and Node.js graphs, with nginx handling server routing, to facilitate a Postgres/Redis backend for the Node.js API server.

v0.3.1#

Added a statement timeout to the pool to prevent long-running queries.

v0.3.0#

Set a default 15-minute statement timeout and implemented monitoring for long-running queries to ensure system efficiency.
Stop propagating run configurable values to the thread configuration, because this can cause issues on subsequent runs if you are specifying a checkpoint_id. This is a slight breaking change in behavior, since the thread value will no longer automatically reflect the unioned configuration of the most recent run. We believe this behavior is more intuitive, however.
Enhanced compatibility with older worker versions by handling event data in channel names within ops.py.

v0.2.137#

Fixed an unbound local error and improved logging for thread interruptions or errors, along with type updates.

v0.2.136#

Added enhanced logging to aid in debugging metaview issues.
Upgraded executor and runtime to the latest version for improved performance and stability.

v0.2.135#

Ensured async coroutines are properly awaited to prevent potential runtime errors.

v0.2.134#

Enhanced search functionality to improve performance by allowing users to select specific columns for query results.

v0.2.133#

Added count endpoints for crons, threads, and assistants to enhance data tracking (#1132).
Improved SSH functionality for better reliability and stability.
Updated @langchain/langgraph-api to version 0.0.59 to fix an invalid state schema issue.

v0.2.132#

Added Go language images to enhance project compatibility and functionality.
Printed internal PIDs for JS workers to facilitate process inspection via SIGUSR1 signal.
Resolved a run_pkey error that occurred when attempting to insert duplicate runs.
Added ty run command and switched to using uuid7 for generating run IDs.
Implemented the initial Golang runtime to expand language support.

v0.2.131#

Added support for object agent spec with descriptions in JS.

v0.2.130#

Added a feature flag (FF_RICH_THREADS=false) to disable thread updates on run creation, reducing lock contention and simplifying thread status handling.
Utilized existing connections for aput and apwrite operations to improve performance.
Improved error handling for decoding issues to enhance data processing reliability.
Excluded headers from logs to improve security while maintaining runtime functionality.
Fixed an error that prevented mapping slots to a single node.
Added debug logs to track node execution in JS deployments for improved issue diagnosis.
Changed the default multitask strategy to enqueue, improving throughput by eliminating the need to fetch inflight runs during new run insertions.
Optimized database operations for Runs.next and Runs.sweep to reduce redundant queries and improve efficiency.
Improved run creation speed by skipping unnecessary inflight runs queries.

v0.2.129#

Stopped passing internal LGP fields to context to prevent breaking type checks.
Exposed content-location headers to ensure correct resumability behavior in the API.

v0.2.128#

Ensured synchronized updates between configurable and context in assistants, preventing setup errors and supporting smoother version transitions.

v0.2.127#

Excluded unrequested stream modes from the resumable stream to optimize functionality.

v0.2.126#

Made access logger headers configurable to enhance logging flexibility.
Debounced the Runs.stats function to reduce the frequency of expensive calls and improve performance.
Introduced debouncing for sweepers to enhance performance and efficiency (#1147).
Acquired a lock for TTL sweeping to prevent database spamming during scale-out operations.

v0.2.125#

Updated tracing context replicas to use the new format, ensuring compatibility.

v0.2.123#

Added an entrypoint to the queue replica for improved deployment management.

v0.2.122#

Utilized persisted interrupt status in join to ensure correct handling of user’s interrupt state after completion.

v0.2.121#

Consolidated events to a single channel to prevent race conditions and optimize startup performance.
Ensured custom lifespans are invoked on queue workers for proper setup, and added tests.

v0.2.120#

Restored the original streaming behavior of runs, ensuring consistent inclusion of interrupt events based on stream_mode settings.
Optimized Runs.next query to reduce average execution time from ~14.43ms to ~2.42ms, improving performance.
Added support for stream mode “tasks” and “checkpoints”, normalized the UI namespace, and upgraded @langchain/langgraph-api for enhanced functionality.

v0.2.117#

Added a composite index on threads for faster searches with owner-based authentication and updated the default sort order to updated_at for improved query performance.

v0.2.116#

Reduced the default number of history checkpoints from 10 to 1 to optimize performance.

v0.2.115#

Optimized cache re-use to enhance application performance and efficiency.

v0.2.113#

Improved thread search pagination by updating response headers with X-Pagination-Total and X-Pagination-Next for better navigation.

v0.2.112#

Ensured sync logging methods are awaited and added a linter to prevent future occurrences.
Fixed an issue where JavaScript tasks were not being populated correctly for JS graphs.

v0.2.111#

Fixed JS graph streaming failure by starting the heartbeat as soon as the connection opens.

v0.2.110#

Added interrupts as default values for join operations while preserving stream behavior.

v0.2.109#

Fixed an issue where config schema was missing when config_type was not set, ensuring more reliable configurations.

v0.2.108#

Prepared for LangGraph v0.6 compatibility with new context API support and bug fixes.

v0.2.107#

Implemented caching for authentication processes to enhance performance and efficiency.
Optimized database performance by merging count and select queries.

v0.2.106#

Made log streams resumable, enhancing reliability and improving user experience when reconnecting.

v0.2.105#

Added a heapdump endpoint to save memory heap information to a file.

v0.2.103#

Used the correct metadata endpoint to resolve issues with data retrieval.

v0.2.102#

Captured interrupt events in the wait method to preserve previous behavior from langgraph 0.5.0.
Added support for SDK structlog in the JavaScript environment for enhanced logging capabilities.

v0.2.101#

Corrected the metadata endpoint for self-hosted deployments.

v0.2.99#

Improved license check by adding an in-memory cache and handling Redis connection errors more effectively.
Reloaded assistants to preserve manually created ones while discarding those removed from the configuration file.
Reverted changes to ensure the UI namespace for gen UI is a valid JavaScript property name.
Ensured that the UI namespace for generated UI is a valid JavaScript property name, improving API compliance.
Enhanced error handling to return a 422 status code for unprocessable entity requests.

v0.2.98#

Added context to langgraph nodes to improve log filtering and trace visibility.

v0.2.97#

Improved interoperability with the ckpt ingestion worker on the main loop to prevent task scheduling issues.
Delayed queue worker startup until after migrations are completed to prevent premature execution.
Enhanced thread state error handling by adding specific metadata and improved response codes for better clarity when state updates fail during creation.
Exposed the interrupt ID when retrieving the thread state to improve API transparency.

v0.2.96#

Added a fallback mechanism for configurable header patterns to handle exclude/include settings more effectively.

v0.2.95#

Avoided setting the future if it is already done to prevent redundant operations.
Resolved compatibility errors in CI by switching from typing.TypedDict to typing_extensions.TypedDict for Python versions below 3.12.

v0.2.94#

Improved performance by omitting pending sends for langgraph versions 0.5 and above.
Improved server startup logs to provide clearer warnings when the DD_API_KEY environment variable is set.

v0.2.93#

Removed the GIN index for run metadata to improve performance.

v0.2.92#

Enabled copying functionality for blobs and checkpoints, improving data management flexibility.

v0.2.91#

Reduced writes to the checkpoint_blobs table by inlining small values (null, numeric, str, etc.). This means we don’t need to store extra values for channels that haven’t been updated.

v0.2.90#

Improve checkpoint writes via node-local background queueing.

v0.2.89#

Decoupled checkpoint writing from thread/run state by removing foreign keys and updated logger to prevent timeout-related failures.

v0.2.88#

Removed the foreign key constraint for thread in the run table to simplify database schema.

v0.2.87#

Added more detailed logs for Redis worker signaling to improve debugging.

v0.2.86#

Honored tool descriptions in the /mcp endpoint to align with expected functionality.

v0.2.85#

Added support for the on_disconnect field to runs/wait and included disconnect logs for better debugging.

v0.2.84#

Removed unnecessary status updates to streamline thread handling and updated version to 0.2.84.

v0.2.83#

Reduced the default time-to-live for resumable streams to 2 minutes.
Enhanced data submission logic to send data to both Beacon and LangSmith instance based on license configuration.
Enabled submission of self-hosted data to a LangSmith instance when the endpoint is configured.

v0.2.82#

Addressed a race condition in background runs by implementing a lock using join, ensuring reliable execution across CTEs.

v0.2.81#

Optimized run streams by reducing initial wait time to improve responsiveness for older or non-existent runs.

v0.2.80#

Corrected parameter passing in the logger.ainfo() API call to resolve a TypeError.

v0.2.79#

Fixed a JsonDecodeError in checkpointing with remote graph by correcting JSON serialization to handle trailing slashes properly.
Introduced a configuration flag to disable webhooks globally across all routes.

v0.2.78#

Added timeout retries to webhook calls to improve reliability.
Added HTTP request metrics, including a request count and latency histogram, for enhanced monitoring capabilities.

v0.2.77#

Added HTTP metrics to improve performance monitoring.
Changed the Redis cache delimiter to reduce conflicts with subgraph message names and updated caching behavior.

v0.2.76#

Updated Redis cache delimiter to prevent conflicts with subgraph messages.

v0.2.74#

Scheduled webhooks in an isolated loop to ensure thread-safe operations and prevent errors with PYTHONASYNCIODEBUG=1.

v0.2.73#

Fixed an infinite frame loop issue and removed the dict_parser due to structlog’s unexpected behavior.
Throw a 409 error on deadlock occurrence during run cancellations to handle lock conflicts gracefully.

v0.2.72#

Ensured compatibility with future langgraph versions.
Implemented a 409 response status to handle deadlock issues during cancellation.

v0.2.71#

Improved logging for better clarity and detail regarding log types.

v0.2.70#

Improved error handling to better distinguish and log TimeoutErrors caused by users from internal run timeouts.

v0.2.69#

Added sorting and pagination to the crons API and updated schema definitions for improved accuracy.

v0.2.66#

Fixed a 404 error when creating multiple runs with the same thread_id using on_not_exist="create".

v0.2.65#

Ensured that only fields from assistant_versions are returned when necessary.
Ensured consistent data types for in-memory and PostgreSQL users, improving internal authentication handling.

v0.2.64#

Added descriptions to version entries for better clarity.

v0.2.62#

Improved user handling for custom authentication in the JS Studio.
Added Prometheus-format run statistics to the metrics endpoint for better monitoring.
Added run statistics in Prometheus format to the metrics endpoint.

v0.2.61#

Set a maximum idle time for Redis connections to prevent unnecessary open connections.

v0.2.60#

Enhanced error logging to include traceback details for dictionary operations.
Added a /metrics endpoint to expose queue worker metrics for monitoring.

v0.2.57#

Removed CancelledError from retriable exceptions to allow local interrupts while maintaining retriability for workers.
Introduced middleware to gracefully shut down the server after completing in-flight requests upon receiving a SIGINT.
Reduced metadata stored in checkpoint to only include necessary information.
Improved error handling in join runs to return error details when present.

v0.2.56#

Improved application stability by adding a handler for SIGTERM signals.

v0.2.55#

Improved the handling of cancellations in the queue entrypoint.
Improved cancellation handling in the queue entry point.

v0.2.54#

Enhanced error message for LuaLock timeout during license validation.
Fixed the $contains filter in custom auth by requiring an explicit ::text cast and updated tests accordingly.
Ensured project and tenant IDs are formatted as UUIDs for consistency.

v0.2.53#

Resolved a timing issue to ensure the queue starts only after the graph is registered.
Improved performance by setting thread and run status in a single query and enhanced error handling during checkpoint writes.
Reduced the default background grace period to 3 minutes.

v0.2.52#

Now logging expected graphs when one is omitted to improve traceability.
Implemented a time-to-live (TTL) feature for resumable streams.
Improved query efficiency and consistency by adding a unique index and optimizing row locking.

v0.2.51#

Handled CancelledError by marking tasks as ready to retry, improving error management in worker processes.
Added LG API version and request ID to metadata and logs for better tracking.
Added LG API version and request ID to metadata and logs to improve traceability.
Improved database performance by creating indexes concurrently.
Ensured postgres write is committed only after the Redis running marker is set to prevent race conditions.
Enhanced query efficiency and reliability by adding a unique index on thread_id/running, optimizing row locks, and ensuring deterministic run selection.
Resolved a race condition by ensuring Postgres updates only occur after the Redis running marker is set.

v0.2.46#

Introduced a new connection for each operation while preserving transaction characteristics in Threads state update() and bulk() commands.

v0.2.45#

Enhanced streaming feature by incorporating tracing contexts.
Removed an unnecessary query from the Crons.search function.
Resolved connection reuse issue when scheduling next run for multiple cron jobs.
Removed an unnecessary query in the Crons.search function to improve efficiency.
Resolved an issue with scheduling the next cron run by improving connection reuse.

v0.2.44#

Enhanced the worker logic to exit the pipeline before continuing when the Redis message limit is reached.
Introduced a ceiling for Redis message size with an option to skip messages larger than 128 MB for improved performance.
Ensured the pipeline always closes properly to prevent resource leaks.

v0.2.43#

Improved performance by omitting logs in metadata calls and ensuring output schema compliance in value streaming.
Ensured the connection is properly closed after use.
Aligned output format to strictly adhere to the specified schema.
Stopped sending internal logs in metadata requests to improve privacy.

v0.2.42#

Added timestamps to track the start and end of a request’s run.
Added tracer information to the configuration settings.
Added support for streaming with tracing contexts.

v0.2.41#

Added locking mechanism to prevent errors in pipelined executions.

Edit this page on GitHub or file an issue.

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.

Link last verified June 7, 2026. View original ↗

Source: LangChain Docs

Link last verified: 2026-03-04