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.
Continue openai-adapters 1.7.0 adds model-capability checks before applying full file diffs and polishes the assistant selector.
└──▷ 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 model capability check before using full file diffs, preventing incompatible models from receiving diff payloads they cannot handle.
›Polishes the 'select assistant' UI experience with visual and interaction improvements.
Continue 1.6.0 adds mercury-coder model support, a new edit completions endpoint, non-streaming API method, and rebrands assistants as agents.
└──▷ 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 overrides to handle /edit/completions requests for the next-edit provider, wiring it to the new edit endpoint.
›Adds a non-streaming method alongside the existing streaming path for OpenAI adapter calls.
›Adds mercury-coder-specific constants and prompt generation templates, replacing the previous nextedit model reference with mercury-coder.
›Rebrands assistants as agents throughout the product UI and internals.
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
CrewAI 0.165.1 adds agent message history to ExternalMemory, auto-injects trigger payloads, and links memory entries to agent IDs in Mem0.
└──▷ GET THIS VERSION
$ git clone --branch 0.165.1 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:$ git checkout 0.165.1
›Includes exchanged agent messages in ExternalMemory metadata for richer memory context.
›Automatically injects crewai_trigger_payload into crew execution context.
›Adds support for agent_id-linked memory entries in Mem0 integration.
└──▷ BREAKING ON UPGRADE
!The internal flag inject_trigger_input is renamed to allow_crewai_trigger_context; any code or config referencing the old name will break.
!The AgentOps integration has been removed; any setup relying on it will no longer function.
CrewAI 0.165.0 enriches ExternalMemory with agent messages, auto-injects trigger payloads, and adds agent_id-linked Mem0 entries.
└──▷ GET THIS VERSION
$ git clone --branch 0.165.0 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:$ git checkout 0.165.0
›Includes exchanged agent messages in ExternalMemory metadata for richer memory context.
›Automatically injects crewai_trigger_payload into crew runs triggered externally.
›Adds support for agent_id-linked memory entries in Mem0 integration.
└──▷ BREAKING ON UPGRADE
!The internal flag inject_trigger_input is renamed to allow_crewai_trigger_context; any configuration or code referencing inject_trigger_input will break.
!The AgentOps integration has been removed; any setup relying on it will stop working.
Haystack v2.17.0 adds image support for 12 model providers, ReasoningContent in ChatMessage, and ByteStream routing in MetadataRouter.
└──▷ GET THIS VERSION
$ git clone --branch v2.17.0 https://github.com/deepset-ai/haystack.git
# already have the repo? check out this version:$ git checkout v2.17.0
└──▷ USE IT
Store model reasoning output in an assistant message for providers that expose chain-of-thought or reasoning traces.
python
from haystack.dataclasses import ChatMessage
msg = ChatMessage.from_assistant(
text="The answer is 42.",
reasoning="First I considered the problem domain, then narrowed down..."
)
print(msg.reasoning)
Override the active tool set at runtime in a previously built pipeline without rebuilding it.
›Adds ReasoningContent as a new content part to ChatMessage, storable via the reasoning parameter in ChatMessage.from_assistant(), enabling assistant messages to carry model reasoning text and metadata.
›Extends SentenceWindowRetriever's source_id_meta_field parameter to accept a list of strings, so only documents matching all specified meta fields are retrieved.
›Adds raise_on_failure parameter to FileTypeRouter (default False); when set to True, always raises FileNotFoundError for non-existent files.
›Extends ToolInvoker.run() to accept a tools list argument that overrides the tools set at construction time, enabling runtime tool switching in pre-built pipelines.
›Adds support for the | union type operator (Python 3.10+) in serialize_type and Pipeline.connect(), alongside existing typing.Union support.
+5 moreshow less
›Expands multimodal image support to Amazon Bedrock, Anthropic, Azure, Google, Hugging Face API, Meta Llama API, Mistral, Nvidia, Ollama, OpenAI, OpenRouter, and STACKIT providers.
›Adds multimodal support to HuggingFaceAPIChatGenerator for vision-language model usage, allowing both text and images to be sent via Hugging Face APIs.
›Extends MetadataRouter to route list[ByteStream] objects in addition to list[Documents].
›Adds serialization/deserialization methods for TextContent and ImageContent parts of ChatMessage.
›Supports subclasses of ChatMessage in Agent state schema validation, checking issubclass(args[0], ChatMessage) instead of requiring exact type equality.
└──▷ BREAKING ON UPGRADE
!MultiFileConverter now outputs a new failed key in its result dictionary containing files that failed to convert; the documents output is only included when at least one file is successfully converted (previously documents could be present but empty).
!HuggingFaceAPIChatGenerator now applies the updated finish_reason mapping consistently regardless of streaming mode: eos_token → stop, stop_sequence → stop, tool calls present → tool_calls. Previously this mapping was only applied when streaming was enabled.
AutoGen 0.7.3 adds GPT-5 model info and extended Pydantic anyOf/oneOf typing support.
└──▷ GET THIS VERSION
$ git clone --branch python-v0.7.3 https://github.com/microsoft/autogen.git
# already have the repo? check out this version:$ git checkout python-v0.7.3
›Adds model info entry for GPT-5, enabling it to be referenced in AutoGen model configurations.
›Extends Pydantic model capability to support anyOf/oneOf item typing for richer schema definitions.
PydanticAI v0.7.3 adds a CLI clipboard command, lets FallbackModel accept string names, and splits Usage into RequestUsage and RunUsage.
└──▷ GET THIS VERSION
$ git clone --branch v0.7.3 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:$ git checkout v0.7.3
└──▷ USE IT
Use a string model name directly in FallbackModel instead of constructing a model object.
python
from pydantic_ai.models.fallback import FallbackModel
model = FallbackModel('openai:gpt-4o', 'anthropic:claude-3-5-sonnet-latest')
›Adds /cp command to the CLI to copy the last response to the clipboard.
›FallbackModel now accepts plain string model names in addition to model objects.
›Moves system_prompt_role from OpenAIModel to OpenAIModelProfile, making it configurable at the profile level.
›Introduces RequestUsage and RunUsage as replacements for the unified Usage class, providing finer-grained usage tracking.
└──▷ BREAKING ON UPGRADE
!Usage is deprecated in favour of RequestUsage and RunUsage; code referencing Usage directly will need to migrate to the appropriate replacement class.
browser-use 0.6.0 drops Playwright in favor of cdp-use/bubus, adds cross-origin iframe support and StreamableHTTP for MCP
└──▷ GET THIS VERSION
$ git clone --branch 0.6.0 https://github.com/browser-use/browser-use.git
# already have the repo? check out this version:$ git checkout 0.6.0
└──▷ TRY IT
Suppress verbose cdp_use debug output in CI by setting the log level to WARNING at the environment level.
$ CDP_USE_LOG_LEVEL=WARNING python my_agent.py
›Sets CDP_USE_LOG_LEVEL environment variable to control cdp_use logging verbosity independently.
›Adds StreamableHTTP transport support to the MCP integration, alongside the existing transport options.
›Replaces Playwright with cdp-use and bubus in BrowserSession, switching to a pure CDP extraction layer for browser automation.
›Adds support for cross-origin iframes, enabling interaction with embedded third-party content.
›Renames ClickEvent(new_tab) parameter to while_holding_ctrl for clarity in click-event semantics.
└──▷ BREAKING ON UPGRADE
!BrowserSession no longer uses Playwright — it now depends on cdp-use and bubus; any code that imported or configured Playwright-specific APIs through BrowserSession will break on upgrade.
!ClickEvent(new_tab) parameter is renamed to while_holding_ctrl; callers passing new_tab by keyword will break.
Arize Phoenix Evals v0.28.0 adds dynamic concurrency control for LLM evaluations.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-evals-v0.28.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-evals-v0.28.0
›Adds dynamic concurrency to LLM evaluation runs, allowing throughput to adjust automatically during execution.
└──▷ BREAKING ON UPGRADE
!Batch evaluation methods have been removed from the evals library.
LanceDB v0.21.4-beta.0 adds train=False and name parameters for index creation.
└──▷ GET THIS VERSION
$ git clone --branch v0.21.4-beta.0 https://github.com/lancedb/lancedb.git
# already have the repo? check out this version:$ git checkout v0.21.4-beta.0
›Adds train=False and name parameters when creating indices, enabling untrained index creation and explicit index naming.
LanceDB python-v0.24.4-beta.0 adds train=False and name parameters for index creation.
└──▷ GET THIS VERSION
$ git clone --branch python-v0.24.4-beta.0 https://github.com/lancedb/lancedb.git
# already have the repo? check out this version:$ git checkout python-v0.24.4-beta.0
›Adds train=False and name parameters when creating indices, allowing users to name indices and skip the training step.
THE AI TOOLCHAIN