The AI Toolchain — issue -365, June 30, 2025

Agno v1.6.4 adds Brightdata web scraping, OpenCV webcam capture, DiscordClient bot integration, and a FileTools search method.

$ git clone --branch v1.6.4 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.6.4

• ›Adds search method to FileTools for searching files within the toolkit.
• ›Adds Brightdata Toolkit with multiple web-based tools including web scraping and data feed capabilities.
• ›Adds OpenCV Video/Image Toolkit with tools for capturing images and video via webcam.
• ›Adds DiscordClient app for connecting an agent or team to Discord as a Discord bot.

Agno v1.6.3 adds store_events to RunResponse, metadata filtering for CSV knowledge bases, and user control flows on the Playground.

$ git clone --branch v1.6.3 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.6.3

• ›Adds store_events parameter to RunResponse/TeamRunResponse to optionally persist all events that occurred during an agent or team run.
• ›Adds metadata filtering support for csv and csv_url knowledge base types.
• ›Adds user control flows support on the Agno Platform Playground.
• ›Shows team member responses during team runs on the Agno Platform Playground.
• ›Shows behind-the-scenes activity during agent and team runs on the Agno Platform Playground.

Agno v1.6.1 adds Nebius embeddings, Firestore memory/storage, async DocumentKnowledgeBase, and enum support in custom tools.

$ git clone --branch v1.6.1 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.6.1

team = Team(
    members=[...],
    stream_member_events=False
)

• ›Adds stream_member_events to team configuration to optionally disable streaming of member events.
• ›Adds agent_name to agent events and team_name to team events in event payloads; adds team_session_id to team-member events.
• ›Adds enum parameter support in custom tools across all models.
• ›Adds async support to DocumentKnowledgeBase.
• ›Adds Nebius as a supported embedding model provider.

Agno v1.6.0 overhauls streaming events for agents, teams, and workflows with granular typed events and member-event propagation.

$ git clone --branch v1.6.0 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.6.0

response = agent.run('Analyze the logs')
if response.status == 'CANCELLED':
    print('Run was cancelled before completion')
elif response.status == 'PAUSED':
    print('Run is awaiting input')

• ›Adds RunResponseContent, RunError, RunCancelled, ToolCallStarted, and ToolCallCompleted event types to agent streaming runs via agent.run(..., stream=True) or agent.arun(..., stream=True).
• ›Adds RunStarted, RunCompleted, ReasoningStarted, ReasoningStep, ReasoningCompleted, MemoryUpdateStarted, and MemoryUpdateCompleted intermediate event types for agents when stream_intermediate_steps=True.
• ›Adds RunResponse.status attribute indicating whether a run response is RUNNING, PAUSED, or CANCELLED.
• ›Adds team-scoped streaming event types — TeamRunResponseContent, TeamRunError, TeamRunCancelled, TeamToolCallStarted, TeamToolCallCompleted — plus intermediate events (TeamRunStarted, TeamRunCompleted, TeamReasoningStarted, TeamReasoningStep, TeamReasoningCompleted, TeamMemoryUpdateStarted, TeamMemoryUpdateCompleted) when stream_intermediate_steps=True.
• ›Teams now propagate and yield streaming events from individual team members as they execute, surfacing member-level activity in the top-level event stream.

Agno v1.5.10 adds Playground file upload, async evals, and an Exa Research tool integration.

$ git clone --branch v1.5.10 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.5.10

• ›Adds research tool to ExaTools, enabling deep research queries against Exa's research API directly from agents.
• ›Adds async support to all evaluations, allowing evals to run non-blocking in async workflows.
• ›Adds file upload support to the Agno Playground, routing PDF, CSV, DOCX, and other files directly to agents/teams or to an attached knowledge base.

Agno v1.5.9 adds AG-UI app, vLLM, LightRAG, 4 new toolkits, PDFBytesKnowledgeBase, and location-aware agents

$ git clone --branch v1.5.9 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.5.9

from agno.knowledge.pdf_bytes import PDFBytesKnowledgeBase
import httpx

pdf_bytes = httpx.get('https://example.com/report.pdf').content
kb = PDFBytesKnowledgeBase(pdf_bytes=pdf_bytes)
kb.load()

from agno.agent import Agent
from agno.models.openai import OpenAIChat

agent = Agent(
    model=OpenAIChat(id='gpt-4o'),
    add_location_to_instructions=True,
)
agent.print_response('What businesses near me are open right now?')

from agno.agent import Agent
from agno.models.openai import OpenAIChat
from agno.tools.serper import SerperTools

agent = Agent(
    model=OpenAIChat(id='gpt-4o'),
    tools=[SerperTools()],
)
agent.print_response('What are the latest CVEs disclosed this week?')

• ›Adds SerperTools toolkit to enable agents to search Google via Serper.
• ›Adds DaytonaTools toolkit to let agents execute code remotely on Daytona sandboxes.
• ›Adds AWSSESTools toolkit to send emails via AWS SES.
• ›Adds PDFBytesKnowledgeBase class to ingest in-memory PDF content via bytes or IO streams instead of file paths.
• ›Adds add_location_to_instructions parameter to automatically detect and inject the agent's current location into the system message.

Agno v1.5.8 adds SlackApp, VisualizationTools, BraveSearch toolkit, and reworks FastAPIApp/WhatsappAPI serving

$ git clone --branch v1.5.8 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.5.8

from agno.agent import Agent
from agno.tools.visualization import VisualizationTools

agent = Agent(
    name="ChartAgent",
    tools=[VisualizationTools()],
)
agent.print_response("Plot a bar chart of monthly sales: Jan=120, Feb=95, Mar=140")

from agno.agent import Agent
from agno.tools.brave_search import BraveSearch

agent = Agent(
    name="WebSearchAgent",
    tools=[BraveSearch()],
)
agent.print_response("What are the latest CVEs disclosed this week?")

• ›Adds SlackApp class to build Slack-connected agents that respond to direct messages, group chats, and automatically create threads for replies.
• ›Adds VisualizationTools toolkit (backed by matplotlib) giving agents the ability to generate graphs.
• ›Adds BraveSearch toolkit so agents can search the web via the Brave Search API.
• ›Adds infer as a parameter to Mem0Tools, exposing inference control in the memory toolkit.
• ›Passes knowledge_filters through when self.add_references=True (traditional RAG path), keeping filter behavior consistent with Agentic RAG.

AutoGPT Platform v0.6.13 adds Google Sign-in, AI/ML API for LLM blocks, VEO3 video generation, and async-first execution.

$ git clone --branch autogpt-platform-beta-v0.6.13 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:
$ git checkout autogpt-platform-beta-v0.6.13

• ›Adds AI/ML API as a supported provider in LLM blocks, expanding model access beyond existing integrations.
• ›Adds VEO3 as a supported model in the AI video generator block.
• ›Adds Google Sign-in support for user authentication.
• ›Makes the execution engine async-first, enabling deeper concurrency for agent graph runs.
• ›Adds depth-first execution ordering to the execution engine, changing how node dependencies are traversed.

AutoGPT Platform v0.6.12 adds an AI Image Editor block, file multipart uploads in web requests, and nested dynamic pin-name support.

$ git clone --branch autogpt-platform-beta-v0.6.12 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:
$ git checkout autogpt-platform-beta-v0.6.12

• ›Adds file multipart upload support to SendWebRequestBlock, enabling blocks to send binary file payloads in web requests.
• ›Adds a new AI Image Editor block powered by Flux Kontext for in-platform image editing.
• ›Adds nested dynamic pin-name support, allowing more flexible dynamic wiring of block inputs and outputs.
• ›Enforces a minimum password length of 12 characters for platform accounts.

CrewAI 0.130.0 adds LiteAgent with Guardrail integration, async tool execution, and multi-org CLI support.

$ git clone --branch 0.130.0 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:
$ git checkout 0.130.0

• ›Introduces LiteAgent with built-in Guardrail integration for lightweight, constrained agent workflows.
• ›Enables async tool execution for more efficient, non-blocking agent workflows.
• ›Adds support for multi-org actions in the CLI.
• ›Upgrades LiteLLM to support the latest OpenAI version.

CrewAI 0.126.0 adds Python 3.13 support, streamable-HTTP MCP transport, prompt/memory transparency, and tool-logging.

$ git clone --branch 0.126.0 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:
$ git checkout 0.126.0

• ›Adds Python 3.13 support.
• ›Adds streamable-HTTP transport support in MCP integration.
• ›Enables tools to be loaded from an Agent repository via their own module.
• ›Persists available tools from a Tool repository across sessions.
• ›Logs tool usage when called by an LLM for observability.

DSPy 2.6.26 adds dspy.Tool as an input field type and dspy.ToolCall as an output field type.

$ git clone --branch 2.6.26 https://github.com/stanfordnlp/dspy.git
# already have the repo? check out this version:
$ git checkout 2.6.26

import dspy

class InvokeTool(dspy.Signature):
    tool: dspy.Tool = dspy.InputField()
    question: str = dspy.InputField()
    tool_call: dspy.ToolCall = dspy.OutputField()

predictor = dspy.Predict(InvokeTool)

• ›Supports dspy.Tool as an input field type and dspy.ToolCall as an output field type, enabling typed tool-calling signatures in DSPy programs.

DSPy 2.6.25 adds CodeAct module, dspy.Audio type, LangChain tool support, and per-module LM history tracking.

$ git clone --branch 2.6.25 https://github.com/stanfordnlp/dspy.git
# already have the repo? check out this version:
$ git checkout 2.6.25

import dspy

class TranscribeAudio(dspy.Signature):
    audio: dspy.Audio = dspy.InputField()
    transcript: str = dspy.OutputField()

predictor = dspy.Predict(TranscribeAudio)
result = predictor(audio=dspy.Audio.from_url("https://example.com/sample.wav"))
print(result.transcript)

import dspy

lm = dspy.LM("openai/gpt-4o")
dspy.configure(lm=lm)

classify = dspy.Predict("text -> label")
classify(text="Suspicious login from unknown IP")

# Inspect history scoped to this module only
print(classify.history)

• ›Adds dspy.Audio as a new built-in field type for passing audio inputs through Signatures.
• ›Adds CodeAct module (dspy.CodeAct) enabling code-execution-based agentic reasoning loops.
• ›Adds LangChain tool support, allowing LangChain tools to be used directly within DSPy modules.
• ›Adds per-module LM history, enabling each module instance to track its own LM call history independently.
• ›Adds a standard base class for creating custom Signature field types, enabling user-defined typed fields in Signatures.

Haystack v2.14.2 adds raise_on_failure to OpenAI document embedders for stricter API error handling.

$ git clone --branch v2.14.2 https://github.com/deepset-ai/haystack.git
# already have the repo? check out this version:
$ git checkout v2.14.2

from haystack.components.embedders import OpenAIDocumentEmbedder

embedder = OpenAIDocumentEmbedder(raise_on_failure=True)

• ›Adds raise_on_failure boolean parameter to OpenAIDocumentEmbedder and AzureOpenAIDocumentEmbedder: when set to True, the component raises an exception on API errors instead of silently logging and continuing; defaults to False to preserve existing behavior.

langchain-openai 0.3.26 adds output format control and automatic response chaining for the Responses API.

$ git clone --branch langchain-openai==0.3.26 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-openai==0.3.26

• ›Adds support for specifying the output format for the Responses API, giving callers control over structured response shapes.
• ›Adds an attribute to always use previous_response_id, enabling automatic response chaining across Responses API calls.

langchain-groq 0.3.3 adds access to reasoning output from Groq models

$ git clone --branch langchain-groq==0.3.3 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-groq==0.3.3

• ›Adds support for accessing reasoning output from Groq models via the langchain-groq integration.
• ›Removes the Python upper bound version constraint for langchain and related libraries, enabling use with newer Python releases.

LangChain 0.3.26 adds pluggable hashing functions for embeddings and Anthropic code execution, MCP connector, and files API support.

$ git clone --branch langchain==0.3.26 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain==0.3.26

• ›Adds Anthropic support for code execution, MCP connector, and files API features.

langchain-openai adds Responses API support to BaseChatOpenAI and AzureChatOpenAI, including streaming.

$ git clone --branch langchain-openai==0.3.24 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-openai==0.3.24

• ›Adds Responses API attributes to BaseChatOpenAI, enabling opt-in routing to the OpenAI Responses API when those attributes are set.
• ›Supports Responses API streaming in AzureChatOpenAI, bringing parity with the standard OpenAI client.

langchain-huggingface 0.3.0 cuts package disk footprint by 95% by making large dependencies optional

$ git clone --branch langchain-huggingface==0.3.0 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-huggingface==0.3.0

• ›Reduces package disk footprint by 95% by making large dependencies (such as transformers) optional — install only what your use case requires.

langchain-tests 0.3.20 adds PDF and audio input support in Chat Completions format and removes Python version upper bound.

$ git clone --branch langchain-tests==0.3.20 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-tests==0.3.20

• ›Supports PDF and audio input in the Chat Completions format for chat model standard tests.
• ›Removes the Python upper bound constraint from langchain and related libraries, enabling use with future Python releases.
• ›Adds benchmark tests to the standard test suite.
• ›Adds a condition gate for the image tool message test to prevent false failures in environments that lack image support.

langchain-anthropic now stores cache TTL details on usage metadata for Anthropic API calls.

$ git clone --branch langchain-anthropic==0.3.15 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-anthropic==0.3.15

• ›Adds cache TTL details to usage metadata returned from Anthropic API calls.

langchain-openai 0.3.19 adds image generation support to the Responses API and caches the httpx client for performance.

$ git clone --branch langchain-openai==0.3.19 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-openai==0.3.19

• ›Adds image generation capability to the OpenAI Responses API integration.
• ›Caches the httpx client to reduce connection overhead across repeated calls.

LangGraph 0.4.10 adds 'tasks' and 'checkpoints' stream modes and numpy/pandas serialization support.

$ git clone --branch 0.4.10 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:
$ git checkout 0.4.10

async for chunk in graph.astream(inputs, stream_mode=["tasks", "checkpoints"]):
    print(chunk)

• ›Introduces tasks and checkpoints stream modes for finer-grained visibility into graph execution.
• ›Supports numpy array serialization in JsonPlusSerializer, enabling numpy data in graph state.
• ›Adds pickle fallback for pandas serialization/deserialization via JsonPlusSerializer.
• ›Allows same-name channels and nodes in StateGraph, removing a previous naming constraint.
• ›Skips saving checkpoints for subgraphs when checkpoint_during=False, reducing unnecessary checkpoint overhead.

langgraph-checkpoint 2.1.0 adds NumPy array and pandas serialization support to JsonPlusSerializer

$ git clone --branch checkpoint==2.1.0 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:
$ git checkout checkpoint==2.1.0

• ›Supports NumPy array serialization in JsonPlusSerializer, enabling checkpoint storage of array-heavy state.
• ›Adds pickle fallback for pandas serialization, allowing DataFrames and Series to round-trip through the checkpoint layer.

LangGraph CLI 0.2.11 adds image_distro config support and warns when distro is not set to Wolfi.

$ git clone --branch cli==0.2.11 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:
$ git checkout cli==0.2.11

• ›Supports image_distro setting in the LangGraph config file for controlling the base image distribution used in Dockerfile generation.
• ›Adds a warning when the image distro is not configured as Wolfi, nudging users toward the recommended distro.

LangGraph 0.4.8 adds NodeBuilder to replace Channel.subscribe_to and flips the default for checkpoint_during.

$ git clone --branch 0.4.8 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:
$ git checkout 0.4.8

• ›Adds NodeBuilder class as the new way to define node subscriptions, replacing Channel.subscribe_to.
• ›Flips the default value for checkpoint_during, changing checkpoint behavior out of the box.
• ›Stream modes messages and custom now respect subgraphs=False, giving finer control over subgraph output filtering.
• ›Requires state_schema in StateGraph.__init__, enforcing explicit schema declaration at graph construction.

Letta 0.7.29 adds configurable batch size and lookback for batch operations.

$ git clone --branch 0.7.29 https://github.com/letta-ai/letta.git
# already have the repo? check out this version:
$ git checkout 0.7.29

• ›Adds configurable batch size and lookback parameters to batch processing.

LlamaIndex v0.12.43 adds ag-ui protocol, openGauss vector store, Hive Intelligence search tool, async MongoDB reader, and mermaid workflow diagrams.

$ git clone --branch v0.12.43 https://github.com/run-llama/llama_index.git
# already have the repo? check out this version:
$ git checkout v0.12.43

from llama_index.utils.workflow import draw_all_possible_flows

draw_all_possible_flows(MyWorkflow, filename="workflow.html")

• ›Adds llama-index-protocols-ag-ui package with ag-ui protocol support for agentic UI integrations.
• ›Adds llama-index-vector-stores-opengauss [0.1.0] with openGauss vector store integration.
• ›Adds llama-index-tools-hive [0.1.0] with a Hive Intelligence search tool.
• ›Adds async driver support via alazy_load_data to llama-index-readers-mongodb.
• ›Adds cache_dir parameter to the Sentence Transformers post-processor in llama-index-postprocessor-sbert-rerank.

LlamaIndex v0.12.42 adds reasoning support for Mistral/Magistral, OpenAI o3-pro, a multimodal OpenAI-like LLM package, figure retrieval, and an ArtifactEditorToolSpec.

$ git clone --branch v0.12.42 https://github.com/run-llama/llama_index.git
# already have the repo? check out this version:
$ git checkout v0.12.42

• ›New llama-index-tools-artifact-editor [0.1.0] package introduces ArtifactEditorToolSpec for editing Pydantic objects as a tool.
• ›New llama-index-multi-modal-llms-openai-like [0.1.0] package adds an OpenAI-compatible multi-modal LLM integration.
• ›Adds reasoning support (including Magistral) to llama-index-llms-mistralai [0.6.0].
• ›Adds day-0 support for OpenAI o3-pro in llama-index-llms-openai [0.4.5].
• ›Adds figure retrieval SDK integration to llama-index-indices-managed-llama-cloud [0.7.7].

LlamaIndex v0.12.41 adds ApertureDB property graph, ElevenLabs voice agents, Ollama thinking, OpenAI JSON Schema output, and Milvus upsert support.

$ git clone --branch v0.12.41 https://github.com/run-llama/llama_index.git
# already have the repo? check out this version:
$ git checkout v0.12.41

• ›Adds MutableMappingKVStore to llama-index-core for easier in-process caching backed by any MutableMapping implementation.
• ›Adds DocumentBlock support to the LiteLLM LLM integration (llama-index-llms-litellm 0.5.1), enabling multimodal document inputs through LiteLLM.
• ›Adds support for Ollama's think feature in llama-index-llms-ollama 0.6.2, exposing model chain-of-thought reasoning.
• ›Adds OpenAI JSON Schema structured output support in llama-index-llms-openai 0.4.4.
• ›Adds log recording during MCP tool calls in llama-index-tools-mcp 0.2.5.

LlamaIndex v0.12.40 adds StopEvent validation, static AWS credentials for Anthropic Bedrock, a Measure Space tool pack, and MCP client header support.

$ git clone --branch v0.12.40 https://github.com/run-llama/llama_index.git
# already have the repo? check out this version:
$ git checkout v0.12.40

from llama_index.tools.mcp import BasicMCPClient

client = BasicMCPClient(
    url="https://my-mcp-server.example.com",
    headers={"Authorization": "Bearer <token>"}
)

• ›Adds header handling to BasicMCPClient in llama-index-tools-mcp, enabling authenticated MCP connections.
• ›New llama-index-tools-measurespace [0.1.0] package adds weather, climate, air quality, and geocoding tools from Measure Space.
• ›Supports passing static AWS credentials to Anthropic Bedrock via llama-index-llms-anthropic.
• ›Enforces StopEvent step validation in llama-index-core workflows so only one step can handle a StopEvent.

AutoGen v0.6.0 adds concurrent GraphFlow agents, a new OpenAIAgent, callable edge conditions, Streamable HTTP MCP, and broader model support.

$ git clone --branch python-v0.6.0 https://github.com/microsoft/autogen.git
# already have the repo? check out this version:
$ git checkout python-v0.6.0

import asyncio

from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.conditions import MaxMessageTermination
from autogen_agentchat.teams import DiGraphBuilder, GraphFlow
from autogen_ext.models.openai import OpenAIChatCompletionClient

async def main():
    model_client = OpenAIChatCompletionClient(model="gpt-4.1-nano")
    agent_a = AssistantAgent("A", model_client=model_client, system_message="You are a helpful assistant.")
    agent_b = AssistantAgent("B", model_client=model_client, system_message="Translate input to Chinese.")
    agent_c = AssistantAgent("C", model_client=model_client, system_message="Translate input to Japanese.")

    builder = DiGraphBuilder()
    builder.add_node(agent_a).add_node(agent_b).add_node(agent_c)
    builder.add_edge(agent_a, agent_b).add_edge(agent_a, agent_c)
    graph = builder.build()

    team = GraphFlow(
        participants=[agent_a, agent_b, agent_c],
        graph=graph,
        termination_condition=MaxMessageTermination(5),
    )

    async for event in team.run_stream(task="Write a short story about a cat."):
        print(event)

asyncio.run(main())

• ›Enables concurrent agent execution in GraphFlow via fan-out-fan-in patterns — select_speaker now returns List[str] | str.
• ›Adds callable (lambda/function) edge conditions for GraphFlow, replacing keyword substring matching.
• ›New OpenAIAgent backed by the OpenAI Responses API.
• ›Supports Streamable HTTP transport for MCP.
• ›Adds tool_call_summary_msg_format_fct parameter to AssistantAgent for custom tool-call summary formatting.

OpenAI Agents SDK v0.0.19 makes Runner an abstract base class, enabling custom runner implementations.

$ git clone --branch v0.0.19 https://github.com/openai/openai-agents-python.git
# already have the repo? check out this version:
$ git checkout v0.0.19

• ›Converts Runner to an abstract base class, allowing practitioners to subclass and implement custom runner logic.

OpenAI Agents SDK v0.0.18 adds REPL support, dynamic prompt templates, and tool_call_id access in tool context.

$ git clone --branch v0.0.18 https://github.com/openai/openai-agents-python.git
# already have the repo? check out this version:
$ git checkout v0.0.18

from agents import Agent, run_demo_loop

agent = Agent(name='Assistant', instructions='You are a helpful assistant.')

import asyncio
asyncio.run(run_demo_loop(agent))

from agents import Agent, RunContextWrapper, function_tool

@function_tool
def my_tool(ctx: RunContextWrapper, query: str) -> str:
    call_id = ctx.tool_call_id
    # use call_id for logging or stateful tracking
    return f'Handled call {call_id}: {query}'

• ›Adds run_demo_loop REPL helper for interactive agent testing sessions.
• ›Adds tool_call_id access via RunContextWrapper so tool functions can read the ID of the current tool call.
• ›Supports dynamic prompt templates through the OpenAI Prompts feature, enabling centrally managed, versioned agent instructions.
• ›Allows arbitrary keyword arguments to be passed through to the underlying model, enabling access to provider-specific parameters not yet explicitly supported.

v0.0.17 adds Portkey AI tracing, RunErrorDetails for max-turns exceptions, and is_enabled on FunctionTool.

$ git clone --branch v0.0.17 https://github.com/openai/openai-agents-python.git
# already have the repo? check out this version:
$ git checkout v0.0.17

from agents import FunctionTool

def lookup_order(order_id: str) -> str:
    return f"Order {order_id}: shipped"

tool = FunctionTool(
    name="lookup_order",
    description="Look up an order by ID",
    params_json_schema={"type": "object", "properties": {"order_id": {"type": "string"}}, "required": ["order_id"]},
    on_invoke_tool=lookup_order,
    is_enabled=False  # disable until user is authenticated
)

• ›Adds is_enabled field to FunctionTool, allowing tools to be conditionally activated or deactivated at runtime.
• ›Adds RunErrorDetails object to the MaxTurnsExceeded exception, giving callers structured context when an agent run hits its turn limit.
• ›Adds Portkey AI as a tracing provider, enabling traces to be sent to the Portkey observability platform.

PydanticAI v0.3.4 adds sensitive-content scrubbing to agent pipelines.

$ git clone --branch v0.3.4 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.3.4

• ›Adds sensitive content scrubbing to redact or sanitize private data within agent interactions.

PydanticAI v0.3.3 adds NativeOutput and PromptedOutput modes and captures more OpenAI-compatible usage fields.

$ git clone --branch v0.3.3 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.3.3

• ›Adds NativeOutput and PromptedOutput output modes alongside the existing ToolOutput mode, giving agents more control over how structured results are produced.
• ›Captures additional usage fields returned by OpenAI-compatible APIs, surfacing richer token and cost details in Usage objects.
• ›Makes Edge hashable, enabling graph edges to be stored in sets and used as dict keys.

PydanticAI v0.3.0 adds ThinkingPart support, parsing provider thinking blocks into a dedicated message part type.

$ git clone --branch v0.3.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.3.0

• ›Adds ThinkingPart as a new message part type: provider-specific <think>...</think> blocks in text responses are now parsed and surfaced as structured ThinkingPart objects rather than raw text.

PydanticAI v0.2.20 adds a process_tool_call hook for MCP servers and RunContext support in history processors.

$ git clone --branch v0.2.20 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.20

• ›Adds process_tool_call hook to MCP servers, enabling interception and modification of tool arguments, metadata, and return values before and after MCP tool execution.
• ›Adds RunContext support to history processors, giving them access to the full run context when processing conversation history.
• ›Adds ModelSettings.timeout enforcement in GoogleModel, so timeout settings are now respected when calling Google models.

PydanticAI v0.2.19 adds history_processors to Agent and surfaces events for unknown tool calls

$ git clone --branch v0.2.19 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.19

from pydantic_ai import Agent
from pydantic_ai.messages import ModelMessage

def redact_secrets(messages: list[ModelMessage]) -> list[ModelMessage]:
    # drop any message whose text contains an API key pattern
    return [m for m in messages if 'sk-' not in str(m)]

agent = Agent('openai:gpt-4o', history_processors=[redact_secrets])
result = agent.run_sync('What did we discuss earlier?')

• ›Adds history_processors parameter to Agent for programmatic pre-processing of message history before each model call.
• ›Yields streaming events for unknown tool calls instead of silently dropping them, enabling downstream handling of unrecognised tool responses.
• ›Makes infer_provider more flexible, accepting a broader range of inputs when resolving provider from a model string.
• ›Ignores dynamic instructions that return an empty string, preventing blank system-prompt entries from being appended to the message list.

PydanticAI v0.2.18 adds MCP Streamable HTTP, OpenAI Responses API vendor ID, and reuses last message when no prompt is given.

$ git clone --branch v0.2.18 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.18

• ›Exposes the OpenAI Responses API response ID as vendor_id on the model response object.
• ›Adds MCP Streamable HTTP transport implementation.
• ›Reuses the last request from message history automatically when no user prompt is provided, enabling continuation flows without re-supplying context.
• ›Switches Gemini inference to use GoogleModel instead of GeminiModel.

PydanticAI v0.2.17 adds token usage to InstrumentedModel, service_tier for OpenAI, custom httpx clients for MCP, and Gemini direct file URL support.

$ git clone --branch v0.2.17 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.17

from pydantic_ai.models.openai import OpenAIModelSettings

settings = OpenAIModelSettings(service_tier='flex')

• ›Adds service_tier field to OpenAIModelSettings to control OpenAI service tier selection.
• ›Adds token usage metrics to InstrumentedModel for observability of model calls.
• ›Allows users to supply a custom httpx.AsyncClient in MCPServerHTTP for full control over HTTP transport.
• ›Supports fileData field (direct file URL) for GeminiModel and GoogleModel, enabling direct URL-based file inputs.
• ›Suppresses inapplicable sampling settings (temperature, top_p) when targeting OpenAI reasoning models.

PydanticAI v0.2.16 adds HerokuProvider, stop_sequences for Google models, and LangChain community tool integration.

$ git clone --branch v0.2.16 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.16

from pydantic_ai import Agent
from pydantic_ai.providers.heroku import HerokuProvider

agent = Agent(provider=HerokuProvider())

• ›Adds HerokuProvider to connect agents to Heroku-hosted models.
• ›Adds stop_sequences parameter support for Google models.
• ›Adds a convenience method to use LangChain community tools directly within PydanticAI agents.
• ›Improves output type inference when callables are provided as output types.

PydanticAI v0.2.13 adds expected-output support to LLMJudge evaluations.

$ git clone --branch v0.2.13 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.2.13

• ›Adds option to pass expected output to LLMJudge, enabling reference-based LLM evaluation scoring.

Semantic Kernel Python 1.34.0 adds Ollama structured outputs, vector stores preview, and KernelArgument merge operators.

$ git clone --branch python-1.34.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout python-1.34.0

from semantic_kernel.kernel_arguments import KernelArguments

base = KernelArguments(city='Seattle', unit='metric')
overrides = KernelArguments(unit='imperial', verbose=True)
merged = base | overrides
# merged: {city: 'Seattle', unit: 'imperial', verbose: True}

• ›Supports | and |= merge operators for KernelArgument, enabling dict-style merging of kernel arguments.
• ›Adds structured output support for Ollama chat completion.
• ›Introduces vector stores preview for Python.
• ›Adds agent response callbacks that provide full context to callers.
• ›Adds pseudo-streaming support via invoke_stream for Copilot-style agents.

Semantic Kernel vectordata-dotnet-9.6.0 adds Ollama ChatClient extensions, AIContextProvider, CopilotStudioAgent, OpenAI Response Agent, hybrid search, and more.

$ git clone --branch vectordata-dotnet-9.6.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout vectordata-dotnet-9.6.0

• ›Adds AIContextProvider support to Semantic Kernel (.NET), enabling dynamic context injection into chat completions via AIContextProvider implementations, with logging now included.
• ›Introduces CopilotStudioAgent (.NET) for agent interactions with Microsoft Copilot Studio.
• ›Adds new OpenAIResponseAgent (.NET) backed by the OpenAI Responses API.
• ›Implements OnnxRuntimeGenAIChatCompletionService on top of OnnxRuntimeGenAIChatClient (.NET), enabling local ONNX model chat completions.
• ›Adds Ollama ChatClient extension methods for the .NET KernelBuilder.

Semantic Kernel .NET 1.56.0 adds OpenAI Response Agent, CopilotStudioAgent, OnnxRuntimeGenAI chat client, and audio/binary chat prompt support.

$ git clone --branch dotnet-1.56.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout dotnet-1.56.0

• ›Introduces CopilotStudioAgent for .NET Agents, enabling integration with Microsoft Copilot Studio.
• ›Adds OpenAIResponseAgent — a new agent type backed by the OpenAI Responses API.
• ›Implements OnnxRuntimeGenAIChatCompletionService on top of OnnxRuntimeGenAIChatClient, bringing local ONNX Runtime GenAI models into the chat completion abstraction.
• ›Adds usage metadata support for ChatClientChatCompletionService adapter, surfacing token/usage telemetry when using IChatClient-backed completions.
• ›Adds support for audio and binary tags in the chat prompt parser, enabling multimodal prompt construction.

Semantic Kernel Python 1.33.0 adds Bing custom search, BinaryContent file handling for OpenAI Responses API, and streaming token usage emission.

$ git clone --branch python-1.33.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout python-1.33.0

• ›Adds file handling support to BinaryContent for the OpenAI Responses API.
• ›Adds Bing custom search tool content support.
• ›Emits token usage data with streaming chat completion agents.
• ›Normalizes MCP function names to allowed tool-calling values for compatibility.
• ›Removes the model info check in Bedrock connectors, broadening model compatibility.

Semantic Kernel Python 1.32.2 adds streaming agent response callbacks and custom httpx client timeout support.

$ git clone --branch python-1.32.2 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout python-1.32.2

• ›Supports custom httpx client timeout when not using a custom client, allowing fine-grained control over request timing.
• ›Adds streaming agent response callback support in agent orchestrations.

Semantic Kernel dotnet-1.55.0 adds Foundry workflow management, ChatHistoryAgentThread, contextual function selection, hybrid search, and MCP SDK update.

$ git clone --branch dotnet-1.55.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout dotnet-1.55.0

• ›Adds ChatHistoryAgentThread to multi-agent support, enabling agents to share and reference a common chat history thread.
• ›Adds Labels field to Gemini request payloads for annotating inference calls.
• ›Adds hybrid search support to the text search store, combining vector and keyword search.
• ›Adds token usage reporting to responses from the Bedrock connector.
• ›Adds logging to AIContextProvider implementations for observability.

AzureAIAgent dependencies now bundled in the base semantic-kernel package

$ git clone --branch python-1.32.1 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout python-1.32.1

• ›Bundles AzureAIAgent required dependencies into the base pip install semantic-kernel package, eliminating the need for a separate extras install.

browser-use 0.3.2 adds a FileSystem tracker for uploads/downloads, a highlight_elements flag, and Gemini 2.5 Flash support.

$ git clone --branch 0.3.2 https://github.com/browser-use/browser-use.git
# already have the repo? check out this version:
$ git checkout 0.3.2

agent = Agent(
    task='Book a flight to NYC',
    llm=llm,
    highlight_elements=False
)

• ›Adds highlight_elements flag to control whether the agent highlights elements on the page during automation.
• ›Introduces a FileSystem feature that tracks all uploads and downloads the agent has access to in a unified manner.
• ›Adds support for gemini-2.5-flash as an available model.
• ›Makes browser launch timeout configurable via Playwright kwargs.
• ›Improves AgentOutput format and reasoning style for better agent state representation.

browser-use 0.3.0 adds an EventBus for queued async agent tasks and automatic retry on mid-action page navigation.

$ git clone --branch 0.3.0 https://github.com/browser-use/browser-use.git
# already have the repo? check out this version:
$ git checkout 0.3.0

• ›Adds an EventBus to the Agent class for queued async task dispatch, enabling event-driven orchestration of browser automation workflows.
• ›Adds automatic retry logic for actions that fail due to page navigation occurring mid-action, reducing brittle task failures in dynamic sites.

browser-use 0.2.6 adds stealth mode, BrowserSession.kill(), new CLI flags, and auto-applied storage state for existing browsers.

$ git clone --branch 0.2.6 https://github.com/browser-use/browser-use.git
# already have the repo? check out this version:
$ git checkout 0.2.6

session = BrowserSession(keep_alive=True, stealth=True)
await session.start()
agent = Agent(task='...', browser_session=session)
await agent.run()
await session.kill()  # force-close even though keep_alive=True

session = BrowserSession(
    cdp_url='ws://localhost:9222',
    storage_state='storage_state.json',  # auto-applied on connect
)
await session.start()

• ›Adds BrowserSession(stealth=True) and BrowserProfile(stealth=True) as a shortcut to run sessions through patchright for bot-detection evasion.
• ›Adds BrowserSession.kill() to force-close a session even when keep_alive=True is set.
• ›Adds --cdp-url, --user-data-dir, and --profile-directory options to the browser-use CLI.
• ›Auto-applies storage_state.json (cookies/localStorage) even when connecting to an already-running browser via CDP.
• ›Every Agent, BrowserSession, and BrowserProfile instance now carries a unique UUID, making them straightforward to persist to a database.

camel-ai v0.2.68 adds RLCards board-game environments and parallel task execution to Workforce.

$ git clone --branch v0.2.68 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.68

• ›Adds RLCards environments for the Project Loong board game, enabling multi-step reinforcement-learning workflows.
• ›Adds parallelization support to Workforce to improve throughput on multi-agent task pipelines.
• ›Updates Workforce with support for the latest Anthropic models.

Camel v0.2.67 adds Workforce shared memory, batch task assignment, KPI metrics, human-in-the-loop, and Qianfan platform integration.

$ git clone --branch v0.2.67 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.67

• ›Adds share_memory support to Workforce, enabling agents within a workforce to share a common memory store.
• ›Adds batch task assignment to Workforce, allowing multiple tasks to be dispatched in a single operation.
• ›Adds KPI metrics collection to Workforce for monitoring and measuring workforce performance.
• ›Adds human-in-the-loop capability to Workforce, letting a human intervene in task processing at runtime.
• ›Updates Workforce task processing to async, unlocking non-blocking multi-agent pipelines.

camel-ai v0.2.66 adds strict-mode JSON schema enforcement for tool definitions

$ git clone --branch v0.2.66 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.66

• ›Sets tool schema to follow strict mode, enforcing stricter JSON schema validation on tool definitions passed to the model

camel-ai v0.2.65 adds a Task Planning Toolkit, O3-pro model support, persistent browser context, stealth mode, and mock website tooling.

$ git clone --branch v0.2.65 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.65

• ›Replaces ChatAgent.single_iteration with max_iteration_ for controlling agent iteration limits.
• ›Adds a new Task Planning Toolkit for structured agent task planning workflows.
• ›Adds support for the O3-pro model.
• ›Supports persistent browser context and stealth mode for browser-based automation.
• ›Adds mock website capability for browser toolkit testing and simulation.

camel-ai v0.2.64 adds Weaviate vector storage, Langfuse integration, MCP export for Workforce, Crynux LLM provider, and two new model supports.

$ git clone --branch v0.2.64 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.64

workforce = Workforce('My Pipeline')
workforce.add_single_agent_worker('Researcher', worker=researcher_agent)
mcp_server = workforce.to_mcp()
mcp_server.run()

• ›Adds to_mcp method to Workforce, enabling Workforce instances to be exported and used as an MCP server.
• ›Integrates Langfuse as an observability/tracing backend for agent runs.
• ›Adds WeaviateVectorStorage as a new vector storage backend.
• ›Adds Crynux as a new LLM provider.
• ›Adds support for gemini-2.5-pro-preview-06-05 model.

camel-ai v0.2.62 adds a PowerPoint toolkit for agent-driven PPTX generation.

$ git clone --branch v0.2.62 https://github.com/camel-ai/camel.git
# already have the repo? check out this version:
$ git checkout v0.2.62

• ›Adds a new pptx toolkit enabling agents to create and manipulate PowerPoint presentations programmatically.

smolagents v1.18.0 adds parallel tool calls, streaming output, and a new ApiWebSearchTool to ToolCallingAgent

$ git clone --branch v1.18.0 https://github.com/huggingface/smolagents.git
# already have the repo? check out this version:
$ git checkout v1.18.0

from smolagents import ApiWebSearchTool, ToolCallingAgent

search_tool = ApiWebSearchTool(headers={"Authorization": "Bearer <token>"})
agent = ToolCallingAgent(tools=[search_tool], model=model)
agent.run("What are the latest CVEs affecting OpenSSH?")

• ›Adds ApiWebSearchTool class for structured web search capabilities via API, with support for custom headers and params.
• ›Adds configurable tool_choice support in prepare_completion_kwargs for fine-grained control over model tool selection.
• ›Enables ToolCallingAgent to execute multiple tool calls in parallel, improving performance on complex multi-tool tasks.
• ›Adds streaming output support to ToolCallingAgent for improved responsiveness during multi-step tool interactions.
• ›Adds support for passing additional params to MLXModel load and tokenizer.apply_chat_template.

Cline v3.18.0 adds a free Gemini CLI provider and optimizes for Claude 4 and Gemini 2.5 model families.

$ git clone --branch v3.18.0 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.18.0

• ›Adds a Gemini CLI provider that uses local Gemini CLI authentication to access Gemini models at no cost.
• ›Optimizes Cline for the Claude 4 model family, enabling improved performance and new capabilities.
• ›Optimizes Cline for the Gemini 2.5 model family.
• ›Updates the default and recommended model to Claude 4 Sonnet.

Cline v3.17.16 adds taskId metadata to LiteLLM API requests for improved request tracing.

$ git clone --branch v3.17.16 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.16

• ›Adds taskId as metadata to LiteLLM API requests, enabling per-task request tracing in LiteLLM logs and dashboards.

Cline v3.17.14-a adds Claude Code and SAP AI Core as API providers, plus configurable terminal and MCP display settings.

$ git clone --branch v3.17.14-a https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.14-a

• ›Adds Claude Code as a new API provider, enabling integration with Anthropic's Claude Code CLI tool and Claude Max Plan.
• ›Adds SAP AI Core as a new API provider with support for Claude and GPT models.
• ›Adds a configurable default terminal profile setting so users can specify which terminal Cline uses.
• ›Adds a terminal output size constraint setting to cap how much terminal output Cline processes.
• ›Adds MCP Rich Display settings to the settings page for persistent configuration.

Cline v3.17.13 adds Gemini thinking UX, Notifications MCP support, and Grok 3 prompt caching indicator.

$ git clone --branch v3.17.13 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.13

• ›Adds Thinking UX for Gemini models, showing visual feedback during model reasoning steps.
• ›Adds support for Notifications MCP integration with Cline.
• ›Adds prompt caching indicator for Grok 3 models.
• ›Sorts MCP marketplace by newest listings by default for easier discovery of recent servers.

Cline v3.17.12 adds free Grok 3 access, collapsible MCP response panels, and smarter file context ordering.

$ git clone --branch v3.17.12 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.12

• ›Adds free Grok 3 model access via the Cline provider — no API key cost.
• ›Adds collapsible MCP response panels to keep conversations focused while preserving access to detailed MCP output.
• ›Prioritizes open editor tabs at the top of the file context menu when using @ mentions.

Cline v3.17.10 adds Qwen 3 thinking mode, streamable MCP servers, new AskSage models, and a terminal reuse toggle.

$ git clone --branch v3.17.10 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.10

• ›Supports Qwen 3 series models with thinking mode options.
• ›Adds new AskSage models: Claude 4 Sonnet, Claude 4 Opus, GPT 4.1, and Gemini 2.5 Pro.
• ›Supports streamable MCP servers.
• ›Adds a VSCode walkthrough to onboard new users.
• ›Improves Ollama model selection with a filterable dropdown replacing radio buttons.

Cline v3.17.9 adds Claude 4 support, CSV/XLSX uploads, stable Grok-3 models, and task timeline scrolling.

$ git clone --branch v3.17.9 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.17.9

• ›Adds experimental support for the Claude 4 model family.
• ›Adds task timeline scrolling for navigating long task histories.
• ›Supports uploading CSV and XLSX files for data analysis and processing.
• ›Adds stable Grok-3 models to the xAI provider: grok-3, grok-3-fast, grok-3-mini, and grok-3-mini-fast.
• ›Adds new models to the Vertex AI provider.

Continue v1.0.12 adds MCP Streamable HTTP, new YAML config keys, OS-aware terminal, and automatic current-file chat context.

$ git clone --branch v1.0.12-vscode https://github.com/continuedev/continue.git
# already have the repo? check out this version:
$ git checkout v1.0.12-vscode

models:
  - name: my-model
    provider: openai
    model: gpt-4o
    maxStopWords: 4
    defaultCompletionOptions:
      stream: true

rules:
  - name: coding-standards
    description: Enforce team coding standards in every session
    alwaysApply: true
    content: Always use snake_case for variable names.

• ›Adds maxStopWords to model config in YAML, letting users cap the number of stop words sent to a model.
• ›Adds stream to defaultCompletionOptions in YAML config, enabling explicit control over streaming behavior per completion.
• ›Adds description field support in markdown YAML front-matter for rules/prompts.
• ›Adds alwaysApply property to rules config for unconditional rule application regardless of context.
• ›Adds MCP Streamable HTTP transport support, expanding how Model Context Protocol servers can be connected.

Continue v1.0.22 adds Next Edit prediction, proxy-aware completions, verbose fetch logging, and Claude 4 Sonnet/Opus model support.

$ git clone --branch v1.0.22-jetbrains https://github.com/continuedev/continue.git
# already have the repo? check out this version:
$ git checkout v1.0.22-jetbrains

$ # Place your secrets in the workspace-scoped location so they take precedence
mkdir -p <workspace>/.continue
echo 'ANTHROPIC_API_KEY=sk-ant-...' > <workspace>/.continue/.env

• ›Secrets are now fetched from <workspace>/.continue/.env before falling back to <workspace>/.env, giving workspace-scoped credentials priority.
• ›Supports caBundlePath as a data URI, enabling inline CA bundle configuration without a file on disk.
• ›Adds verbose logging option for the custom fetch layer to aid in diagnosing network and proxy issues.
• ›Supports completions through a proxy, extending LLM reachability in firewalled or enterprise environments.
• ›Adds claude-4-sonnet and claude-opus (Opus) to the built-in LLM info registry.

Goose v1.0.29 adds a system prompt parameter to run, a /clear command, alphabetized extensions, and a confirmation dialog for unsaved changes.

$ git clone --branch v1.0.29 https://github.com/block/goose.git
# already have the repo? check out this version:
$ git checkout v1.0.29

$ goose run --system "You are a security-focused code reviewer. Flag any use of eval() or exec()." --recipe review.yaml

$ /clear

• ›Adds --system (system prompt) parameter to the goose run command, enabling custom system-level instructions per run.
• ›Adds /clear command to clear the Goose context mid-session.
• ›Alphabetizes extensions listing in the Goose CLI for easier navigation.
• ›Adds confirmation dialog for unsaved changes when closing the extension modal in the UI.
• ›Adds a Help & Feedback section in App Settings.

Goose v1.0.28 adds SageMaker TGI support, an LLM tool router, recipe settings, and desktop auto-update.

$ git clone --branch v1.0.28 https://github.com/block/goose.git
# already have the repo? check out this version:
$ git checkout v1.0.28

• ›Adds SageMaker TGI as a supported provider backend.
• ›New LLM tool selector for the tool router, enabling AI-driven routing of tool calls.
• ›Tool router now includes extension names in vector DB and search tool metadata.
• ›Goose recipes now support configurable settings.
• ›Adds auto-update functionality to the Goose desktop app, gated by an UPDATES_ENABLED flag.

Goose v1.0.26 adds session export, a temporal scheduler, lead/worker models, Snowflake provider, and Gemini 2.5 support.

$ git clone --branch v1.0.26 https://github.com/block/goose.git
# already have the repo? check out this version:
$ git checkout v1.0.26

$ goose config set GOOSE_CONTEXT_STRATEGY summarize

$ GOOSE_RECIPE_PATH=/team/recipes goose run --recipe my-recipe

• ›New export CLI command converts sessions to Markdown for sharing or archiving.
• ›Adds a temporal scheduler with schedule creation, editing, pause/unpause, live tracking, and task cancellation for scheduled jobs.
• ›New lead/worker model architecture enables multi-model agent workflows.
• ›New goose web command provides a local web-based terminal alternative.
• ›Adds Snowflake as a new LLM provider.

OpenCode v0.1.158 adds experimental hooks, auto-formatting for Elixir and Go, and a scroll-to-last-message button in the web UI.

$ git clone --branch v0.1.158 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.158

• ›Adds experimental hooks feature for extending OpenCode behavior via shell scripts.
• ›Adds auto-formatting support for Elixir files.
• ›Adds auto-formatting support for Go files.
• ›Adds a scroll-to-last-message button in the web UI for quick navigation to the latest output.
• ›Lazy-loads formatters to improve startup performance.

OpenCode v0.1.153 adds Elixir file formatting support and lazy-loads all formatters for faster startup.

$ git clone --branch v0.1.153 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.153

• ›Adds Elixir file formatting support.
• ›Lazy-loads formatters to improve startup performance.

OpenCode v0.1.146 adds LSP diagnostics in the TUI, a default system theme, and expanded theme options.

$ git clone --branch v0.1.146 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.146

• ›Displays LSP diagnostics inline for edit and write tool operations in the TUI.
• ›Adds a default system theme that follows the OS/terminal color scheme.
• ›Expands the theme library with additional built-in TUI themes.
• ›Adds output length error reporting to surface truncation and limit issues explicitly.

OpenCode v0.1.122 adds smarter GitHub Copilot header editing to reduce rate limiting and optimistic user message rendering in the TUI.

$ git clone --branch v0.1.122 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.122

• ›Optimistically renders user messages in the TUI for a more responsive feel.

OpenCode v0.1.120 adds GitHub Copilot OAuth authentication and combines bash tool stdout/stderr output.

$ git clone --branch v0.1.120 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.120

• ›Adds GitHub Copilot OAuth authentication flow for seamless login without manual token management.
• ›Combines stdout and stderr into a single stream in bash tool output for unified command result visibility.

OpenCode v0.1.118 adds a Matrix-inspired theme.

$ git clone --branch v0.1.118 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.118

• ›New Matrix-inspired theme available in the theme dialog.

OpenCode v0.1.113 adds a live theme switcher plus six built-in themes including Nord, Catppuccin, and Gruvbox.

$ git clone --branch v0.1.113 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.113

• ›Adds an in-TUI theme switcher with live preview so you can cycle themes without restarting.
• ›Supports custom themes, letting you define your own color scheme in config.
• ›Adds global config support for session context handling.

OpenCode v0.1.87 overhauls config with a merged global+project JSON format and shows provider next to model in the TUI.

$ git clone --branch v0.1.87 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.87

• ›Adds global config support at ~/.config/opencode/config.json, automatically merged with any project-level config.
• ›Displays the provider name next to the model in the TUI for at-a-glance context.

OpenCode v0.1.61 restructures custom provider config with a new npm field for package loading.

$ git clone --branch v0.1.61 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.1.61

• ›Adds npm field to custom provider config entries to explicitly specify which npm package to load.
• ›Adds validation on provider ID and improved error messages in opencode auth login.

OpenHands 0.46 adds custom model names, vi-mode CLI keybindings, configurable safety settings, and system prompt overrides.

$ git clone --branch 0.46.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.46.0

$ CLI_VI_MODE=true python -m openhands.core.cli

• ›Supports custom/arbitrary model names during CLI model selection, beyond the preset list.
• ›Adds optional vi-style keybindings (j/k navigation) for CLI confirmation prompts, enabled via the CLI_VI_MODE environment variable.
• ›Adds customizable safety settings for Mistral AI and Gemini models.
• ›Enables overriding the hardcoded system prompt for full prompt control.
• ›Allows users to submit feedback on agent performance even when the agent hit an error or is awaiting user input.

OpenHands 0.45 adds Kubernetes runtime, Bitbucket integration, file/image uploads, per-conversation budgets, and Japanese UI.

$ git clone --branch 0.45.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.45.0

• ›Adds Kubernetes Runtime support as a new sandbox execution backend.
• ›Supports file and image uploads directly within conversations.
• ›Adds Bitbucket integration for local usage alongside existing GitHub support.
• ›Enables per-conversation budget configuration to cap LLM spending.
• ›Adds Japanese language support to the UI.

OpenHands 0.44 adds draft PR/MR control via Git MCP, Slack integration in settings, and SWE-bench-Live evaluation support.

$ git clone --branch 0.44.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.44.0

• ›Agents can now control whether PRs/MRs opened via Git MCP are created as drafts or ready-for-review.
• ›Adds Slack integration installable directly from the Integrations tab in the Cloud OpenHands Settings page.
• ›Supports evaluation runs on SWE-bench-Live for more realistic, up-to-date issue-resolving benchmarks.
• ›Dev container networking now works without host network mode, improving security and portability of the dev setup.

OpenHands 0.43 adds Slack integration (beta), a Microagents UI, GitLab lower-scoped token support, and a code review microagent example.

$ git clone --branch 0.43.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.43.0

• ›New Slack integration (beta) with Cloud OpenHands for triggering and interacting with agents via Slack.
• ›New Microagents UI in the conversation context menu lets users see what specialized knowledge microagents are available in their current conversation.
• ›New example code review microagent provides a ready-made template for automated code review workflows.
• ›Supports lower-scoped GitLab tokens, enabling more granular permission control when connecting GitLab repositories.
• ›Adds JSON serialization for array and object parameters when converting tools, broadening tool-call compatibility.

OpenHands 0.42.0 adds API endpoints to control conversation lifecycle and lets users update their email address.

$ git clone --branch 0.42.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.42.0

• ›New API endpoints to explicitly start and stop conversations programmatically.
• ›New VSCODE_IN_NEW_TAB feature flag to open the VS Code editor in a new browser tab.

OpenHands 0.41.0 adds Cloud GitLab Resolver, streamable HTTP MCP support, and native Windows runtime.

$ git clone --branch 0.41.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.41.0

• ›Adds Cloud GitLab Resolver, enabling automated issue/PR resolution against cloud-hosted GitLab repositories.
• ›Implements streamable HTTP MCP transport, supporting MCP servers served over HTTP streams.
• ›Adds native Windows support without requiring WSL, broadening local deployment options.
• ›Improves CLI mode setup flow to guide users through configuration when settings are missing.

OpenHands 0.40.0 adds native PR/MR creation via MCP, LocAgent integration, and a first-class search API.

$ git clone --branch 0.40.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.40.0

• ›Enables native PR/MR opening on GitHub and GitLab via in-house MCP integration.
• ›Incorporates LocAgent into OpenHands for improved code localization capabilities.
• ›Adds first-class search API support to OpenHands.
• ›Adds Interactive SWE-Bench benchmark for evaluating agent performance.
• ›Improves MCP tool usage visualization in the UI.

Zed v0.192.5 adds inline color previews, Vercel AI provider, GitHub Copilot enterprise support, and Gemini 2.5 models.

$ git clone --branch v0.192.5 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.192.5

{ "minimap": { "display_in": "active_editor" } }

• ›Adds { "minimap": { "display_in": "active_editor" } } to settings.json to show the minimap only on the currently active editor.
• ›Implements dap.args setting to pass custom arguments to a debug adapter binary, configurable in .zed/settings.json.
• ›Adds project panel: collapse all entries action to collapse all Project Panel entries without requiring panel focus.
• ›Adds git: open modified files command to open all git-modified files at once.
• ›Adds Vercel as a language model provider in the AI Agent.

Zed v0.191.8 enables saving changes directly from the zed --diff view.

$ git clone --branch v0.191.8 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.191.8

• ›Enables saving changes directly within the zed --diff view.

Zed v0.191.6 lets you change the OpenAI API base URL directly from the Agent UI.

$ git clone --branch v0.191.6 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.191.6

• ›Adds UI control to override the OpenAI API base URL in the Agent panel, enabling use of compatible third-party or self-hosted endpoints.

Zed v0.191.5 adds a debugger onboarding modal and surfaces package.json tasks as debuggable scenarios.

$ git clone --branch v0.191.5 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.191.5

• ›Debugger: New session modal now surfaces tasks from package.json as ready-to-launch debuggable scenarios.
• ›Debugger: Adds an onboarding modal to guide users through initial debugger setup.

Zed v0.191.4 ships a native debugger, zed --diff A B, per-thread agent profiles, and image/thinking support for more AI providers.

$ git clone --branch v0.191.4 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.191.4

$ zed --diff before.yaml after.yaml

{
  "project_panel": {
    "hide_root": true
  }
}

$ export GH_COPILOT_TOKEN=<your-oauth-token>
zed .

• ›Adds zed --diff A B CLI flag to open a native diff view between any two files.
• ›Adds git::PushTo action to select which remote to push to.
• ›Adds git::FetchFrom action to fetch from a single remote.
• ›Adds hide_root config key under project_panel to hide the root entry when only one folder is open.
• ›Adds drag_and_drop_selection config key (set to false to disable drag-and-drop text selection).

Zed v0.190.4 adds Vim mode in the agent panel, OpenRouter support, Python venv auto-config, a Cursor keymap, and FreeBSD SSH remotes.

$ git clone --branch v0.190.4 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.190.4

// In Zed settings (settings.json)
{
  "close_on_file_delete": true
}

// In .zed/tasks.json
[
  {
    "label": "Run tests in file directory",
    "command": "pytest",
    "args": ["$ZED_RELATIVE_DIR"],
    "cwd": "$ZED_WORKTREE_ROOT"
  }
]

• ›Adds close_on_file_delete setting (off by default) to automatically close open editor tabs when the underlying file is deleted from disk.
• ›Adds ZED_RELATIVE_DIR task variable providing the path to the current file's directory relative to the worktree root.
• ›Adds AGENT.md and AGENTS.md as recognized rules file names for the agent panel.
• ›Adds thinking mode for custom Google models in the agent panel, with configurable token budget.
• ›Adds disabled_globs setting path expansion to handle ~ in Edit Prediction glob patterns.

Zed v0.189.5 adds Cursor settings import, an element inspector, inline code action indicators, and major agent/AI improvements.

$ git clone --branch v0.189.5 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.189.5

{
  "agent": {
    "default_view": "text_thread"
  }
}

{
  "inline_code_actions": false
}

• ›Adds dev::ToggleInspector command (debug builds only) to open a pane for inspecting element info and live-modifying styles.
• ›Adds zed: import cursor settings command palette action to import settings from Cursor.
• ›Adds inline_code_actions setting (set to false to disable) to control the new inline code action indicator shown at the start of each row.
• ›Adds minimap.thumb.background and minimap.thumb.border theme keys to customize minimap thumb color and border.
• ›Adds agent.default_view config key (values: thread or text_thread) to choose whether the agent or text thread panel opens by default.

Jan v0.6.0 adds MCP server support, multi-assistant management, API key auth, Vulkan toggle, download resume, fuzzy search, and a Tauri build option.

$ git clone --branch v0.6.0 https://github.com/janhq/jan.git
# already have the repo? check out this version:
$ git checkout v0.6.0

• ›Adds API Key setting to the Jan local API server, enabling authenticated access to the built-in OpenAI-compatible endpoint.
• ›Adds a Vulkan toggle in settings to enable GPU acceleration on supported hardware.
• ›Adds support for resuming interrupted model downloads via a new download manager for the llama.cpp extension.
• ›Adds fuzzy search integration into the model dropdown for faster model selection.
• ›Adds a refresh button to reload the model list for remote providers.

KoboldCpp v1.94.2 adds Chroma image generation, PhotoMaker face cloning, and several new CLI flags for GPU and embedding control.

$ git clone --branch v1.94.2 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:
$ git checkout v1.94.2

$ koboldcpp --model main-model.gguf --embeddingsmodel embed-model.gguf --embeddingsmaxctx 512

• ›Adds --sdphotomaker flag to load PhotoMaker for face cloning alongside any SDXL-based model, available in text2img, img2img, and inpaint modes via the SDUI at http://localhost:5001/sdui.
• ›Adds --sdclampedsoft flag for soft total resolution clamping (e.g. a value of 640 allows 640x640, 512x768, and 768x512 images), combinable with the existing --sdclamped hard-clamp flag.
• ›Adds --sdtiledvae flag to specify a resolution threshold beyond which VAE tiling is applied, replacing the previous --sdnotile.
• ›Adds --embeddingsmaxctx flag to cap the max context length for embedding models, reducing memory usage.
• ›Adds --embeddingsgpu flag to offload embedding model layers to GPU.

KoboldCpp v1.93.2 adds Windows Shell integration for .gguf files, in-memory save/load session states, model unloading via admin API, and a new --embeddingsmaxctx flag.

$ git clone --branch v1.93.2 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:
$ git checkout v1.93.2

$ koboldcpp --model mymodel.gguf --embeddingsmaxctx 2048

$ koboldcpp --model mymodel.gguf --singleinstance

• ›Adds --embeddingsmaxctx CLI option to control the maximum context size for the embeddings endpoint.
• ›Adds --singleinstance flag awareness for Windows Shell integration: only instances started with this flag are replaced when a .gguf file is double-clicked to open in KoboldCpp.
• ›Adds add_sd_step_limit adapter flag to cap the maximum number of steps in image generation.
• ›Adds /api/extra/shutdown API endpoint (localhost-only) to power Windows Shell integration, allowing a running instance to be replaced when a new .gguf is launched.
• ›Adds Windows Shell integration allowing .gguf files to be associated with KoboldCpp; double-clicking a .gguf opens it directly, replacing any existing local instance on the same port. Installable/uninstallable from the 'Extras' tab.

LocalAI v3.1 adds Gemma 3n support, GPU-aware meta-packages in the backend gallery, and a leaner container image layout.

$ git clone --branch v3.1.0 https://github.com/mudler/LocalAI.git
# already have the repo? check out this version:
$ git checkout v3.1.0

$ local-ai run gemma-3n-e4b-it

• ›Adds local-ai run gemma-3n-e2b-it and local-ai run gemma-3n-e4b-it commands to run Google Gemma 3n models (text generation) locally.
• ›Introduces meta-packages to the backend gallery so installing a backend (e.g. vllm) automatically selects the correct GPU variant without manual version picking.
• ›Moves default model and backend paths in container images to /models/ and /backends/ respectively.

LocalAI 3.0 ships a Backend Gallery (OCI-based, API-driven), Realtime WebSocket API, audio/PDF upload in the UI, llama.cpp reranking, and 50+ new models.

$ git clone --branch v3.0.0 https://github.com/mudler/LocalAI.git
# already have the repo? check out this version:
$ git checkout v3.0.0

$ docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gpu-nvidia-cuda-12

$ docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-aio-cpu

• ›Adds a Backend Gallery that lets you install and remove backends at runtime or startup via API or the WebUI, backed by standard OCI images — replaces the old -extra image approach.
• ›Adds Backend Gallery download progress display in the WebUI so you can track backend installations in real time.
• ›Adds Realtime WebSocket API with OpenAI-compatible streaming, suitable for chat apps and agents.
• ›Adds llama.cpp reranking support.
• ›Adds enhanced multimodal support to llama.cpp via libmtmd.

SGLang v0.4.7 adds pipeline parallelism, PD disaggregation with NIXL backend, Qwen3/Kimi-VL/InternVL3/MiMo model support, CUDA graph for LoRA, and full Blackwell GPU support.

$ git clone --branch v0.4.7 https://github.com/sgl-project/sglang.git
# already have the repo? check out this version:
$ git checkout v0.4.7

$ curl http://localhost:30000/v1/chat/completions -H 'Content-Type: application/json' -d '{"model": "Qwen/Qwen3-8B", "messages": [{"role": "user", "content": "Solve step by step: 12 * 47"}], "chat_template_kwargs": {"enable_thinking": true}}'

$ curl http://localhost:30000/v1/chat/completions -H 'Content-Type: application/json' -d '{"model": "meta-llama/Llama-4-Scout", "messages": [{"role": "user", "content": "Summarize this document."}], "max_completion_tokens": 512}'

• ›Adds chat_template_kwargs.enable_thinking to enable thinking mode via the chat template, allowing reasoning/tool-call workflows to be toggled per request.
• ›Supports max_completion_tokens parameter for OpenAI Chat Completions API, aligning with the OpenAI spec.
• ›Adds pipeline parallelism support, enabling multi-node inference pipelines across GPU clusters.
• ›Adds NIXL backend for PD (prefill-decode) disaggregation, now supporting Prefill TP and Decode TP+DP configurations.
• ›Supports expert parallelism (EP) MoE for Qwen3, extending large-scale EP to the Qwen3 model family.

oobabooga textgen v3.5 adds persistent UI settings, .docx attachments, Qwen3 presets, and RTX 50XX CUDA 12.8 support.

$ git clone --branch v3.5 https://github.com/oobabooga/textgen.git
# already have the repo? check out this version:
$ git checkout v3.5

• ›Adds CUDA 12.8 installation option supporting RTX 50XX NVIDIA Blackwell GPUs via ExLlamaV2/V3 and Transformers backends.
• ›Uses user_data/cache/gradio for Gradio temporary files instead of the system temporary folder.
• ›Adds Qwen3 presets ('Thinking' and 'No Thinking'); sets Qwen3 - Thinking as the new default preset.
• ›Adds support for .docx file attachments.
• ›Adds an option (configurable in the Session tab) to automatically convert long pasted text into an attachment; disabled by default.

Ollama v0.9.3 adds support for Google's Gemma 3n models, optimized for laptops, tablets, and phones.

$ git clone --branch v0.9.3 https://github.com/ollama/ollama.git
# already have the repo? check out this version:
$ git checkout v0.9.3

$ ollama run gemma3n:e2b

$ ollama run gemma3n:e4b

• ›Supports Gemma 3n models (e2b and e4b variants), designed for efficient on-device execution across laptops, tablets, and phones with training data covering 140+ spoken languages.

Ollama v0.9.1 adds tool calling for DeepSeek-R1 671B and Magistral, plus a redesigned macOS/Windows preview app.

$ git clone --branch v0.9.1 https://github.com/ollama/ollama.git
# already have the repo? check out this version:
$ git checkout v0.9.1

$ ollama pull deepseek-r1:671b

• ›Adds tool calling support for DeepSeek-R1-2508 (671B) and Magistral models.
• ›Supports disabling thinking mode on Magistral (with recommended system prompt change).
• ›New preview macOS and Windows desktop apps with network exposure, local browser access, and configurable model directory.
• ›macOS app rebuilt as a native application for smaller footprint and faster startup.
• ›Enables exposing Ollama on the network so other devices (or remote users) can reach a central Ollama host.

Arize Phoenix v11.1.0 adds imagePullSecrets support to the Helm chart.

$ git clone --branch arize-phoenix-v11.1.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v11.1.0

• ›Adds imagePullSecrets support to the Helm chart, enabling deployments that pull images from private registries.

Phoenix v11 ships cost tracking for LLM spans, including model cost calculations and updated OpenAI reasoning model support.

$ git clone --branch arize-phoenix-v11.0.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v11.0.0

• ›Adds model cost tracking to LLM spans, calculating token-level costs across supported models.
• ›Adds updated cost definitions for OpenAI reasoning models.

arize-phoenix-otel 0.12.0 adds support for passing tracer provider arguments.

$ git clone --branch arize-phoenix-otel-v0.12.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-otel-v0.12.0

• ›Supports passing tracer provider arguments when initializing the OpenTelemetry integration in arize-phoenix-otel.

Arize Phoenix 10.15.0 adds a Bedrock playground client for interactive model testing.

$ git clone --branch arize-phoenix-v10.15.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v10.15.0

• ›Adds a Bedrock playground client, enabling interactive testing of Amazon Bedrock models directly within the Phoenix UI.

Phoenix now falls back to OTEL_EXPORTER_OTLP_ENDPOINT when PHOENIX_COLLECTOR_ENDPOINT is not set.

$ git clone --branch arize-phoenix-v10.14.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v10.14.0

$ export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317
# PHOENIX_COLLECTOR_ENDPOINT is not set; Phoenix now picks up the OTel standard var automatically

• ›Reads OTEL_EXPORTER_OTLP_ENDPOINT as a fallback endpoint when PHOENIX_COLLECTOR_ENDPOINT is not defined, enabling standard OpenTelemetry environment variable conventions to work without Phoenix-specific configuration.

Arize Phoenix client v1.11.0 adds log_spans, dataset methods, logout, and OTEL endpoint fallback.

$ git clone --branch arize-phoenix-client-v1.11.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-client-v1.11.0

• ›Adds log_spans to the Phoenix client and REST API, enabling programmatic span ingestion.
• ›Adds dataset methods to the Phoenix client for managing datasets directly via the library.
• ›Adds logout support to the auth layer of the Phoenix client.
• ›Falls back to reading OTEL_EXPORTER_OTLP_ENDPOINT when PHOENIX_COLLECTOR_ENDPOINT is not set, easing OTEL-standard deployments.

arize-phoenix-otel v0.11.0 adds logout support, Phoenix Cloud spaces, and fallback to OTEL_EXPORTER_OTLP_ENDPOINT.

$ git clone --branch arize-phoenix-otel-v0.11.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-otel-v0.11.0

$ export OTEL_EXPORTER_OTLP_ENDPOINT=https://your-otel-collector:4318
python your_instrumented_app.py

• ›Falls back to the standard OTEL_EXPORTER_OTLP_ENDPOINT environment variable for the collector endpoint when PHOENIX_COLLECTOR_ENDPOINT is not set.
• ›Adds logout capability to the auth flow.
• ›Enables Phoenix Cloud spaces support.

Phoenix v10.13.0 adds FullStory integration and enhanced Helm service configurability.

$ git clone --branch arize-phoenix-v10.13.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v10.13.0

• ›Adds FullStory session-replay integration to the Phoenix UI.
• ›Enhances Helm chart service configurability with improved configuration options and documentation.