LangChain

This release of langchain-core includes a minor bug fix related to handling empty tool chunk IDs during merging.

This release focuses on improving the Anthropic integration by adding Bedrock support and a User-Agent header, alongside several fixes related to model IDs and response handling.

This release focuses on bug fixes in the core library, improved error handling, and the addition (and subsequent reversion) of the ChatAnthropicBedrock wrapper. Performance improvements were also made by deferring specific langsmith imports.

This release focuses primarily on bug fixes within the core library, addressing issues related to tool call merging, recursion errors, and LangSmith tracing parameter handling. Dependency groups were also updated.

Version 1.1.1 of langchain-text-splitters focuses on bug fixes, including resolving an SSRF vulnerability and mutation issues in splitters, alongside adding support for model_kwargs in SentenceTransformersTokenTextSplitter.

This release bumps several internal dependency groups, most notably updating langsmith to version 0.6.3, and releases version 1.1.5 of langchain-tests.

This release introduces support for automatic server-side compaction in OpenAI and adds the OpenRouter provider package. It also includes several bug fixes related to model properties, error handling, and dependency version bumps.

This release primarily involves bumping the core version and silencing a warning within the langchain-openrouter integration.

This release updates the Anthropic integration to support effort="max" and introduces a new ContextOverflowError for handling context limits across Anthropic and OpenAI integrations.

This release introduces a new ContextOverflowError for handling context limits in Anthropic and OpenAI integrations, and includes a fix for sanitizing URLs during image token counting.

This patch release for langchain-openai (1.1.8) primarily includes a bug fix related to detecting codex models for API response preferences and various dependency and chore updates.

This release primarily focuses on adding standard tests for sandbox providers and minor maintenance chores within the standard-tests package.

This release for langchain-anthropic primarily focuses on bug fixes and feature enhancements, including support for output_config and automatic compaction (Opus 4.6). It also includes dependency updates.

Version 1.1.2 of langchain-groq introduces support for LangChain image types and includes a core serialization fix. Dependencies were also updated.

This release includes an initial version and a dependency upgrade for urllib3 to version 2.6.3.

Version 1.2.2 of langchain-xai primarily addresses a bug related to streaming API routing for chat completions and responses.

This release introduces the new `langchain-openrouter` provider package and expands the documentation for `get_lc_namespace`.

This release primarily contains a bug fix related to setting the text attribute on ChatGeneration objects.

This release primarily focuses on bug fixes, including sanitizing URLs for OpenAI token counting and improving exception handling in the tracer.

This release primarily focuses on dependency updates and includes a fix for token counting issues with partial message sequences. An internal variable name was also updated.

This release introduces a new `ContextOverflowError` for better error handling in LLM integrations and enhances token counting to include tool schemas.

LangChain version 1.2.9 introduces support for state updates via `wrap_model_call` and threading context in agent flows, alongside several bug fixes related to schema normalization and token counting.

This release focuses on improving approximate token counting by allowing scaling based on reported usage and adjusting the capping mechanism. It also reverts a previous change regarding hex color regex precompilation.

This release focuses on minor dependency upgrades, linting, and introduces `ToolCallRequest` to middleware exports, alongside a fix for agent factory name mismatches.

This release focuses on numerous bug fixes across the core library, including improvements to token counting with multimodal support and adding an XML format option for buffer string generation. Dependency updates and various documentation enhancements were also included.

LangChain version 1.2.7 introduces dynamic tool registration through middleware and includes minor fixes to the summarization prompt and system message grammar.

This release primarily addresses a bug fix related to the signature and configuration invocation of the SummarizationMiddleware.

This release updates the summarization prompt and fixes an issue by adding metadata configuration to the summarization model invocation.

This release focuses heavily on internal type checking improvements across many test suites and introduces metadata for agent names. A key functional change is adding the `state` field to `_ModelRequestOverrides`.

This release updates the core library to use uuid7 for generating run identifiers across several components.

This release focuses on improving summarization logic by leveraging usage metadata and fixing tool call pairings, alongside a provider mapping fix for Azure OpenAI.

LangChain 1.2.2 is a patch release focused on fixing parallel tool usage in planning middleware and improving internal test type safety.

This release focuses on bug fixes for the OpenAI integration, specifically improving error handling for structured output refusals and refining token counting logic for function calls.

This release focuses on bug fixes for message summarization and tool schemas, adds Google GenAI support to embeddings, and introduces several linting and type-checking improvements.

Version 1.3.1 is a maintenance release that addresses a security vulnerability (CVE-2025-68664) and fixes several bugs related to cache control and serialization.

Version 1.2.1 of langchain-xai fixes a bug in token usage reporting by ensuring reasoning tokens are included in the total output count.

This release focuses on internal performance optimizations, improved type safety, and bug fixes for ChatPromptTemplate and CallbackManager. It also updates internal tooling and documentation to reflect beta statuses and Pydantic v2 compatibility.

This release updates langchain-xai to 1.2.0 and langchain-openai to 1.1.6, focusing on streaming improvements, token counting fixes for GPT-5, and core serialization patches.

This release includes a critical serialization fix for langchain-core and internal code style improvements across standard-tests and text-splitters.

This release updates langchain-classic to 1.0.1, featuring enhanced model initialization for Anthropic and Google GenAI, a transition to uuid7 for run IDs, and the removal of the Tigris provider.

This release includes a fix for a serialization issue in langchain-core.

This release focuses on bug fixes for tool decorators and serialization, while introducing PEP 702 support for deprecations and automated tool call tracking.

This release introduces usage metadata tracking in LangChainTracer and fixes a trace persistence issue for iterator-based inputs.

This release introduces a fix for handling unknown blocks in OpenAI message conversion and adds infrastructure checks for lockfile consistency.

Version 1.1.6 updates the maximum input token limits for the gpt-5 series models.

Version 1.1.5 of langchain-openai fixes an issue with chunk_position by delegating its setting to langchain-core.

Version 1.1.1 of langchain-tests introduces more descriptive error messages for streaming cases to improve developer experience during testing.

Version 1.2.2 is a maintenance release focusing on Python 3.14 compatibility and internal lockfile updates.

Version 1.1.4 reverts a model selection change for flex usage and fixes documentation links for structured output.

This release introduces a strict mode for structured outputs in ProviderStrategy and adds an extras field to BaseTool, alongside fixes for multithreading race conditions and HuggingFace model initialization.

This patch release fixes a bug in tool call parsing when handling null arguments and improves type hinting for ToolCallChunk.

Release 1.1.0 introduces R language support for text splitting and transitions to uuid7 for run IDs, while ensuring compatibility with Python 3.14.

This release introduces support for the gpt-5.2-pro responses API and includes bug fixes for image resizing and API response handling.

Release 1.2.0 focuses on fixing and improving the init_chat_model integration for HuggingFace backends and updating core dependencies.

Version 1.1.1 of langchain-mistralai fixes a bug regarding null content in tool call responses and updates internal core dependencies.

This release introduces the ability to override tool_choice in structured output and adds parameter filtering for tool binding in the Groq integration.

Version 1.1.0 introduces invocation model overrides and stricter type checking while removing the deprecated has_tool_choice property and improving documentation clarity.

This release fixes a bug regarding unsupported arguments in ChatOllama and adds unit tests for reasoning parameters. It also includes documentation improvements and cross-references for tool conversion functions.

Version 1.3.0 introduces Model Context Protocol (MCP) support, automatic beta header management for computer use, and enhanced tool capabilities including tool search and TypedDict support.

This release introduces an extras field to BaseTool for enhanced provider metadata and improves type flexibility for OpenAI tool conversions.

This release introduces a new Search API for Chroma and resolves critical bugs related to Pydantic validation and OpenCLIP image embedding regressions.

This release fixes parameter handling for temperature and tools in ChatOpenAI and introduces a strict flag for structured outputs.

This release focuses on fixing dependency constraints for jinja2 and resolving issues related to partial function usage in retrievers.

This release introduces support for Upstage models in the chat model initializer and adds an agent name field to AIMessage, alongside typing improvements in core message utilities.

Version 1.1.1 of langchain-openai addresses a bug in the Responses API when mixing tool types and includes minor documentation improvements.

This release introduces Google Maps grounding support and improves asynchronous task stability. It also deprecates the py_anext utility and refines documentation and typing for message utilities.

This release introduces uuid7 for run IDs and provides several bug fixes for tool calling, schema handling, and output parsing formatting.

This patch release simplifies the internal logic for summarization cutoffs.

This release introduces UUIDv7 for run IDs and improves the accuracy of the SummarizationMiddleware by adjusting token counts based on the specific model used.

LangChain 1.1 introduces ModelRetryMiddleware, enhanced SummarizationMiddleware with context window support, and SystemMessage support in agent creation. It also includes critical fixes for tool call ordering and Anthropic tool execution.

This release introduces distributed model profile data across packages and includes various documentation updates and code style refinements.

This release introduces a dedicated output parser for Perplexity reasoning models and expands usage metadata tracking. It also includes internal improvements to model profile data distribution and documentation updates.

This release focuses on distributing model profile data across packages and fixing documentation syntax errors.

This release focuses on distributing model-profile data across packages and fixing validation logic for GPT-5 temperature settings, alongside various documentation and style improvements.

This release focuses on distributing model profile data across packages and includes several documentation improvements and link fixes.

This release introduces distributed model profiles and fixes a bug regarding parameter passing in the ChatHuggingFace class.

This release introduces distributed model profile data and deprecates the use of setattr on ModelCallRequest, alongside documentation syntax fixes.

Release 1.1.0 of langchain-fireworks introduces distributed model profiles and updates the suite of tested models while fixing documentation syntax.

This release introduces native structured output and strict tool calling for Anthropic, alongside various documentation syntax fixes and linting improvements across the codebase.

Version 0.0.5 introduces data distribution across packages for langchain-model-profiles and includes documentation updates.

This release introduces model profiles to facilitate data distribution across packages within the LangChain ecosystem.

This release addresses a security vulnerability (GHSA-6qv9-48xg-fc7f) by tightening validation for f-string templates and restricting functionality in jinja2 and mustache templates.

This release focuses on security hardening by restricting template engine functionality and improving input variable validation for f-string templates.

This is a minor patch release for langchain, incrementing the version from 1.0.7 to 1.0.8.

This release introduces proxy support for Mermaid PNG rendering and enables tool runtime injection for custom argument schemas.

This release introduces native structured output and strict tool calling support for Anthropic models, alongside code execution capabilities and various bug fixes for tool nodes and async hooks.

Version 1.0.3 introduces handling for incomplete streaming events and enforces token limits for embeddings, alongside several bug fixes for API compatibility.

This is a minor release for langchain version 1.0.7, primarily serving as a version bump from 1.0.6.

This release includes cherry-picked bug fixes to improve the integration between the core langchain package and the langchain-anthropic partner package.

This release includes cherry-picked bug fixes for the LangChain and Anthropic integration along with improvements to middleware docstrings.

This release focuses on bug fixes regarding exception handling and reverts a recent change to secret mapping in the load function. It also includes documentation improvements for token counting methods.

Release 1.0.1 introduces prompt caching token usage details and fixes issues with empty tool call arguments and core dependency versioning.

This release introduces support for strict structured output in the DeepSeek integration and includes various documentation and metadata updates.

Minor release for langchain-nomic updating metadata and extending compatibility to Python 3.14.

This release introduces support for the code_execution_20250825 feature in the Anthropic integration and includes minor documentation updates.

Version 0.0.4 of langchain-model-profiles includes general code cleanup and styling improvements.

LangChain 1.0.5 focuses on stability by reverting SystemMessage support in create_agent, increasing the default recursion limit, and updating core dependencies.

This release introduces subgraph drawing support via pygraphviz and includes several bug fixes related to schema generation, async callback context preservation, and prompt formatting.