The daily firehose — everything the toolchain shipped today, already filtered.
// HOW THIS ISSUE IS MADE
We read every release from the 174 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.
An open-source AI agent that brings the power of Gemini directly into your terminal.
gemini-cli v0.8.0-nightly adds tab-completion for extension names in /extensions update.
└──▷ GET THIS VERSION
$ git clone --branch v0.8.0-nightly.20250930.ae51bbda https://github.com/google-gemini/gemini-cli.git
# already have the repo? check out this version:$ git checkout v0.8.0-nightly.20250930.ae51bbda
›Adds extension name auto-complete to the /extensions update command, reducing manual lookup of exact extension names.
langchain-anthropic v1.0.0a2 adds memory/context management, web fetch beta, server tool call/result types, dynamic Max Tokens mapping, cache_control kwarg, and more.
└──▷ GET THIS VERSION
$ git clone --branch langchain-anthropic==1.0.0a2 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-anthropic==1.0.0a2
└──▷ USE IT
Enable parallel tool calls to let Claude invoke multiple tools simultaneously in one turn.
python
from langchain_anthropic import ChatAnthropic
from langchain_core.tools import tool
@tool
def get_weather(city: str) -> str:
"""Get weather for a city."""
return f"Sunny in {city}"
llm = ChatAnthropic(model="claude-opus-4-5", parallel_tool_calls=True)
llm_with_tools = llm.bind_tools([get_weather])
response = llm_with_tools.invoke("What's the weather in Paris and Tokyo?")
›Adds cache_control as a passable kwarg on ChatAnthropic invocations for fine-grained cache control.
›Adds dynamic mapping of Max Tokens for Anthropic models, automatically selecting appropriate limits per model.
›Adds support for memory and context management features in ChatAnthropic.
›Adds server tool call and result types (v1) for use with the Anthropic Messages API.
›Adds web fetch beta support, enabling ChatAnthropic to fetch web content as part of tool use.
+10 moreshow less
›Adds support for code execution, MCP connector, and files API features.
›Adds support for built-in tools via ChatAnthropic.
›Adds parallel_tool_calls support on ChatAnthropic.
›Adds support for passing URLs directly to ChatAnthropic as multimodal content.
›Adds citations support in streaming responses and across multi-turn conversations.
langchain-openai v1.0.0a3 adds server tool call types, PDF URL support, web_search tool, max_tokens for AzureChatOpenAI, and removes bind_functions.
└──▷ GET THIS VERSION
$ git clone --branch langchain-openai==1.0.0a3 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-openai==1.0.0a3
└──▷ USE IT
Cap output tokens when using Azure OpenAI to control cost and latency.
python
from langchain_openai import AzureChatOpenAI
llm = AzureChatOpenAI(
azure_deployment="gpt-4o",
azure_endpoint="https://<your-resource>.openai.azure.com/",
api_version="2024-02-01",
max_tokens=512,
)
response = llm.invoke("Summarise this incident report in one paragraph.")
›Adds max_tokens parameter to AzureChatOpenAI for controlling output token limits.
›Adds web_search to the OpenAI tools list, enabling built-in web search as a callable tool.
›Supports PDFs passed via URL in the standard content format for multimodal chat inputs.
›Introduces server tool call and result types (feat: (v1) server tool call and result types) for the v1 API surface.
›Removes bind_functions from ChatOpenAI/AzureChatOpenAI and moves tool_calls out of additional_kwargs in the v1 interface.
+7 moreshow less
›Drops Python 3.9 support in the v1 package; minimum supported version is now Python 3.10.
›Updates default output_version in the v1 OpenAI integration.
›Adds standard content IDs, translators, and normalization for cross-provider message compatibility.
›Officially supports verbosity parameter in ChatOpenAI for controlling output detail level.
›Supports minimal output mode alongside verbosity in ChatOpenAI.
›Supports custom tools in ChatOpenAI via the custom tools feature.
›Allows overriding ls_model_name from kwargs for LangSmith tracing.
└──▷ BREAKING ON UPGRADE
!bind_functions is deleted from the OpenAI chat model classes in v1; callers must migrate to bind_tools.
!tool_calls is removed from additional_kwargs in v1; code reading additional_kwargs['tool_calls'] will find it missing.
!Python 3.9 is no longer supported; the package requires Python 3.10 or later.
langchain-core v1.0.0a5 adds server tool types, AWS Bedrock content blocks, reasoning content parsing, and expanded OpenAI tool support
└──▷ GET THIS VERSION
$ git clone --branch langchain-core==1.0.0a5 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-core==1.0.0a5
└──▷ USE IT
Filter OpenAI data content blocks from a message using the now-public is_openai_data_block helper.
python
from langchain_core.messages.content import is_openai_data_block
blocks = [b for b in message.content if is_openai_data_block(b)]
Strip PostgreSQL NUL bytes from text before inserting into a vector store to avoid DataError.
python
from langchain_core.utils import sanitize_for_postgres
clean_text = sanitize_for_postgres(raw_text)
vectorstore.add_texts([clean_text])
›Adds reasoning_content parsing from additional_kwargs in message conversion, enabling structured access to model reasoning traces.
›Adds support for the reasoning type in convert_to_openai_messages for models that emit reasoning content.
›Adds web_search to the OpenAI tools list, expanding the set of built-in tool types recognized by the framework.
›Adds is_openai_data_block as a public API with filtering support for inspecting and filtering OpenAI data content blocks.
›Adds id field to Document objects passed to the filter function in InMemoryVectorStore similarity search.
+12 moreshow less
›Adds server tool call and result types (v1) for representing tool interactions in a standardized server-side format.
›Adds standard content block support for AWS Bedrock, including document content blocks in msg_content_output.
›Adds support for PromptTemplates with formats other than f-string.
›Adds sanitize_for_postgres utility to strip PostgreSQL NUL bytes that cause DataError.
›Adds an option to make deserialization more permissive.
›Allows overriding ls_model_name from kwargs when tracing.
AutoGen 0.7.5 adds Anthropic thinking mode, linear memory for RedisMemory, and reasoning_effort for GPT-5 models.
└──▷ GET THIS VERSION
$ git clone --branch python-v0.7.5 https://github.com/microsoft/autogen.git
# already have the repo? check out this version:$ git checkout python-v0.7.5
└──▷ USE IT
Control reasoning depth when calling GPT-5 models to balance latency and answer quality.
python
from autogen_ext.models.openai import OpenAIChatCompletionClient
client = OpenAIChatCompletionClient(
model="gpt-5",
reasoning_effort="high",
)
›Adds thinking mode support for the Anthropic client, enabling extended reasoning with Claude models.
›Supports linear memory storage in RedisMemory, giving practitioners an alternative memory retrieval strategy.
›Adds reasoning_effort parameter support for OpenAI GPT-5 models.
›Adds security warnings and defaults to DockerCommandLineCodeExecutor for safer code execution.
Arize Phoenix v12.1.0 adds dataset label GraphQL mutations, Claude Sonnet 4.5 support, and a Helm chart migration.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v12.1.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v12.1.0
›Adds dataset label GraphQL operations (create, delete, read) for managing labels on datasets via the API.
›Supports Claude Sonnet 4.5 as a model option in the playground.
›Migrates the Helm chart from the Bitnami chart to the groundhog2k chart.
LanceDB v0.22.2-beta.1 adds bitmap index support for large-string, binary, and large-binary types, plus remote connection testing.
└──▷ GET THIS VERSION
$ git clone --branch v0.22.2-beta.1 https://github.com/lancedb/lancedb.git
# already have the repo? check out this version:$ git checkout v0.22.2-beta.1
›Extends bitmap indexes to cover large-string, binary, large-binary, and bitmap column types.
›Adds support for test_remote_connections to validate remote database connectivity.
LanceDB python-v0.25.2-beta.1 adds bitmap index support for large-string, binary, and large-binary types.
└──▷ GET THIS VERSION
$ git clone --branch python-v0.25.2-beta.1 https://github.com/lancedb/lancedb.git
# already have the repo? check out this version:$ git checkout python-v0.25.2-beta.1
›Allows bitmap indexes to be created on large-string, binary, large-binary, and bitmap column types.
›Adds support for test_remote_connections to validate remote database connectivity.
THE AI TOOLCHAIN