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.
Agno v1.8.1 adds Neo4j graph tools, OpenAI reasoning summaries, and timezone support for Teams.
└──▷ GET THIS VERSION
$ git clone --branch v1.8.1 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:$ git checkout v1.8.1
└──▷ USE IT
Query and traverse a Neo4j graph database from an Agno agent to investigate relationships between entities.
python
from agno.agent import Agent
from agno.tools.neo4j import Neo4jTools
agent = Agent(
tools=[Neo4jTools(uri='bolt://localhost:7687', user='neo4j', password='<password>')],
show_tool_calls=True,
)
agent.print_response('Find all nodes connected to the user with id 42')
Get a readable reasoning summary from an OpenAI reasoning model response instead of raw chain-of-thought tokens.
python
from agno.agent import Agent
from agno.models.openai import OpenAIResponses
agent = Agent(
model=OpenAIResponses(id='o3', summary='auto'),
markdown=True,
)
agent.print_response('Explain the steps to assess a phishing email')
›Adds Neo4jTools class to explore and manipulate graphs in a Neo4j database from within agents.
›Adds reasoning summary support to the OpenAIResponses class for OpenAI reasoning models.
›Adds timezone_identifier field to Team, bringing it to parity with the existing Agent functionality.
›Allows OpenAIChat to accept both httpx.Client and httpx.AsyncClient for custom HTTP client configuration.
AutoGPT Platform v0.6.25 adds DataForSEO, BaaS, Stagehand, Ideogram V3, and Discord OAuth2 blocks plus admin agent rejection.
└──▷ GET THIS VERSION
$ git clone --branch autogpt-platform-beta-v0.6.25 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:$ git checkout autogpt-platform-beta-v0.6.25
›Adds DiscordGetCurrentUserBlock for retrieving OAuth2 user details from Discord within agent workflows.
›Adds DataForSEO keyword research blocks for SEO data retrieval in agent pipelines.
langchain-openai v1.0.0a1 adds Responses API support, standard content blocks, custom tools, verbosity control, and drops Python 3.9
└──▷ GET THIS VERSION
$ git clone --branch langchain-openai==1.0.0a1 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-openai==1.0.0a1
└──▷ USE IT
Enforce parallel tool call behavior explicitly when binding tools to a model.
python
from langchain_openai import ChatOpenAI
from langchain_core.tools import tool
@tool
def lookup_cve(cve_id: str) -> str:
"""Fetch details for a CVE."""
...
llm = ChatOpenAI(model="gpt-4o")
llm_with_tools = llm.bind_tools([lookup_cve], parallel_tool_calls=False)
›Adds verbosity parameter to ChatOpenAI for controlling response verbosity level.
›Adds minimal mode alongside verbosity for trimmed response output.
›Adds parallel_tool_calls as an explicit keyword argument to bind_tools.
›Adds Responses API attributes to BaseChatOpenAI, enabling routing to the OpenAI Responses API when relevant attributes are set.
›Adds previous_response_id attribute to always chain Responses API calls.
+25 moreshow less
›Supports output format specification for the Responses API.
›Supports Responses API streaming in AzureChatOpenAI.
›Adds image generation capability to the Responses API.
›Supports built-in code interpreter and remote MCP tools via the Responses API.
›Supports multi-turn computer use via the Responses API.
›Supports structured output and tools via the Responses API.
›Supports streaming reasoning summaries from OpenAI reasoning models.
›Supports streaming token counts in AzureChatOpenAI.
›Adds token counting support for o-series models in ChatOpenAI.
›Adds explicit service_tier attribute and propagates service_tier to response metadata.
›Supports standard multi-modal content blocks (audio, PDF, image) in convert_to_openai_messages.
›Adds custom tools support to ChatOpenAI.
›Adds encoding_model attribute to allow explicit specification of the tokenization model.
›Supports runtime kwargs in OpenAIEmbeddings.
›Removes tool_calls from additional_kwargs and deletes bind_functions in v1.0 cleanup.
›Updates BaseChatModel return type to AIMessage.
›Introduces standard content IDs, translators, and normalization across content block types.
›Adds max_retries parameter to ChatOpenAI for handling 503 capacity errors.
›Uses max_completion_tokens in place of max_tokens for compatible models.
›Updates system role to developer for o-series models.
›Enables streaming for o1 models.
›Supports json_schema response format with streaming.
›Supports serialization of Pydantic models in messages.
›Caches the httpx client for improved connection reuse.
›Runs _tokenize in a background thread during async embedding invocations.
└──▷ BREAKING ON UPGRADE
!Python 3.9 is no longer supported; minimum supported version is now Python 3.10.
!bind_functions is removed from ChatOpenAI.
!tool_calls is removed from additional_kwargs in ChatOpenAI responses.
langchain-anthropic 1.0.0a1 adds standard content blocks, cache_control kwargs, parallel tool calls, web search, code execution, MCP connector, files API, and dynamic Max Tokens mapping.
└──▷ GET THIS VERSION
$ git clone --branch langchain-anthropic==1.0.0a1 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-anthropic==1.0.0a1
└──▷ USE IT
Run multiple tool calls in parallel to speed up agent steps that can fan out independently.
python
from langchain_anthropic import ChatAnthropic
llm = ChatAnthropic(model='claude-3-5-sonnet-20241022')
llm_with_tools = llm.bind_tools([search_tool, calculator_tool], parallel_tool_calls=True)
response = llm_with_tools.invoke('What is the weather in Paris and what is 42 * 17?')
›Adds cache_control as a passthrough kwarg on ChatAnthropic for fine-grained prompt caching control.
›Supports parallel_tool_calls parameter on ChatAnthropic to enable or disable parallel tool execution.
›Adds dynamic mapping of Max Tokens for Anthropic models, automatically selecting appropriate token limits per model.
Semantic Kernel .NET adds 'minimal' reasoning effort support and stable APIs for OpenAI/Azure connectors.
└──▷ GET THIS VERSION
$ git clone --branch dotnet-1.64.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:$ git checkout dotnet-1.64.0
›Adds support for 'minimal' reasoning effort level for Azure and OpenAI connectors, giving practitioners a third tier alongside existing effort options.
›Promotes previously experimental items to stable (removes experimental flags) in the .NET SDK.
›Adds input and output attributes to invoke_agent OpenTelemetry spans for richer observability of agent calls.
└──▷ BREAKING ON UPGRADE
!Updated encoding logic in prompt templates changes how template arguments are encoded; existing prompts relying on prior encoding behavior may render differently — see the Semantic Kernel blog post for migration details.
Continue 1.11.0 adds --mcp and --model CLI flags, multi-edit and single search-and-replace tools, a Unified Terminal, and a FIM/next-edit toggle.
└──▷ GET THIS VERSION
$ git clone --branch @continuedev/[email protected] https://github.com/continuedev/continue.git
# already have the repo? check out this version:$ git checkout @continuedev/[email protected]
└──▷ TRY IT
Run Continue against a specific model and MCP server from the command line, useful in CI or scripted workflows.
$ continue --model gpt-4o --mcp my-mcp-server
›Adds --mcp and --model flags to the CLI for specifying MCP server and model at invocation time.
›Introduces multi-edit and single search-and-replace tools for applying targeted code changes across files.
›Adds UI for diff tools, surfacing multi-edit and single-replace operations in the editor interface.
›Adds single-replace to the default tools set.
›Adds a user-friendly toggle between FIM (fill-in-the-middle) and next-edit autocomplete modes.
Continue 1.11.0 adds a user-friendly toggle between FIM and next-edit autocomplete modes.
└──▷ GET THIS VERSION
$ git clone --branch @continuedev/[email protected] https://github.com/continuedev/continue.git
# already have the repo? check out this version:$ git checkout @continuedev/[email protected]
›Adds a user-friendly toggle to switch between FIM (fill-in-the-middle) and next-edit autocomplete modes.
Qwen Code v0.0.9-nightly.4 adds subcommands to switch between project and global memory operations.
└──▷ GET THIS VERSION
$ git clone --branch v0.0.9-nightly.4 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:$ git checkout v0.0.9-nightly.4
›Adds subcommands to switch between project-level and global memory operations, replacing a single undifferentiated memory command.
›Adds "auto_indent": false to settings (globally or per language) to control automatic indentation per language.
›Adds status_bar.cursor_position_button setting (boolean, default true) to show or hide the cursor position button in the status bar.
›Adds ZED_BUILD_REMOTE_SERVER=nomusl environment variable to opt out of musl builds when building the remote server.
›Adds [f/]f Vim keybindings to navigate to the next collaborator.
›Helix: binds alt-s to editor: split selection into lines.
+14 moreshow less
›Launches Agent Client Protocol (ACP), an open standard for integrating external agents with Zed, with Google's Gemini CLI as the first integration.
›Adds grok-code-fast-1 model to xAI's list of available models in the Agent.
›Adds file status indicators in multibuffer headers.
›Adds a context menu for buffer titles.
›Adds a button in the Keymap Editor to open the keymap JSON file, with right-click shortcuts to open default Zed and Vim keymaps.
›Improves Python virtual environment propagation into LSP configuration for more reliable Go to Definition and Find All References.
›Improves project panel empty state to include git clone action and quick local folder opening.
›Text entered in a new untitled buffer is now used as the default filename suggestion when saving.
›Adds multiple selection support to the editor: unwrap syntax node action in Vim mode.
›Helix: adds support for yanking a single character under cursor with y when no active selection exists.
›Agent fetch tool now requires user confirmation before executing.
›Agent font size now inherits UI font size by default when not explicitly set in settings.
›Agent automatically selects a language model provider when no user-set provider is configured.
›Adds drop-target highlighting when dragging a tab between two other tabs.
└──▷ BREAKING ON UPGRADE
!The /docs slash command has been removed.
!The Zed Plex Sans and Zed Plex Mono font names are replaced by IBM Plex Sans and Lilex respectively; old names remain as aliases for backward compatibility.
!.ZedSans and .ZedMono are introduced as new aliases for the default fonts, currently resolving to IBM Plex Sans and Lilex.
!The Services menu has been removed on non-macOS systems.
Phoenix v11.29.0 improves the experiments compare list page styling.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v11.29.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v11.29.0
›Improves styling of the experiments compare list page UI.
Milvus Go client v2.6.0 adds search iterator v2, function reranker, IvfRabitQ & MinHashLSH indexes, and per-field analyzer support.
└──▷ GET THIS VERSION
$ git clone --branch client/v2.6.0 https://github.com/milvus-io/milvus.git
# already have the repo? check out this version:$ git checkout client/v2.6.0
›Adds search iterator v2 with a limit parameter and static param validation for the search iterator.
›Supports IvfRabitQ and MinHashLSH index types.
›Supports a function reranker for search results.
›Enables running an analyzer scoped to a specific collection field.
THE AI TOOLCHAIN