1 Release published by 1 person

• v1.0.1

  published Jul 18, 2025

14 Pull requests merged by 8 people

• fix: include agent trace into tool for agent as tools

  #526 merged Jul 23, 2025

• feat(a2a): support mounts for containerized deployments

  #524 merged Jul 23, 2025

• Automatically flatten nested tool collections

  #508 merged Jul 23, 2025

• ci: enable integ tests for anthropic, cohere, mistral, openai, writer

  #510 merged Jul 22, 2025

• docs: add hot reloading documentation for load_tools_from_directory

  #517 merged Jul 22, 2025

• Doc fixes: suppressing some typos in various texts

  #487 merged Jul 22, 2025

• Don't re-run workflows on un/approvals

  #516 merged Jul 22, 2025

• ci: expose STRANDS_TEST_API_KEYS_SECRET_NAME to integration tests

  #513 merged Jul 22, 2025

• deps(a2a): address interface changes and bump min version

  #515 merged Jul 22, 2025

• Update to use dedicated github logo

  #505 merged Jul 21, 2025

• fix(telemetry): group traces when using agent as tool in an agent

  #493 merged Jul 18, 2025

• fix(agent): prevent JSON serialization errors with non-serializable direct tool parameters

  #498 merged Jul 18, 2025

• docs: Update README.md with Writer

  #474 merged Jul 18, 2025

• fix: enable parallel execution in graph workflow

  #485 merged Jul 18, 2025

4 Pull requests opened by 4 people

• feature(graph): Allow cyclic graphs

  #497 opened Jul 18, 2025

• Add VPC endpoint support to BedrockModel class - Add optional endpoin…

  #502 opened Jul 20, 2025

• Update pyproject.toml

  #504 opened Jul 21, 2025

• fix! Litellm handle non streaming response fix for issue #477

  #512 opened Jul 22, 2025

23 Issues closed by 9 people

• [BUG] @tool decorator fails to inject parent agent when called via agent.tool.toolname()

  #518 closed Jul 23, 2025

• [BUG] Strands import tools from `tools/tools.py` even when no tools are imported.

  #376 closed Jul 22, 2025

• [BUG] Class Method with @tool decorator is not getting accepted as Tool

  #199 closed Jul 22, 2025

• [FEATURE] Ability to filter callback / iterator messages

  #250 closed Jul 22, 2025

• [BUG] Previous wrongly forced selection of us-west-2 region by agents not fixed by Merge #254 and passage to v1.0.0

  #479 closed Jul 22, 2025

• [FEATURE] Disabling tool message in agent.messages when tools run in event loop

  #266 closed Jul 22, 2025

• [BUG] Tool response truncation occurs silently

  #167 closed Jul 22, 2025

• [FEATURE] Strands Agents should provide a built-in intent confidence

  #211 closed Jul 22, 2025

• [BUG] SlidingWindowConversationManager does not work as (I) expect

  #465 closed Jul 22, 2025

• [BUG] OTEL context not passed from Agent to Agent

  #489 closed Jul 21, 2025

• [ANNOUNCEMENT] Upcoming Breaking Changes

  #253 closed Jul 21, 2025

• [BUG] Agent object not JSON serializable when passed to direct tool calls

  #350 closed Jul 18, 2025

• [BUG] Token counting inconsistency with image_reader tool - metrics don't reflect actual token usage

  #460 closed Jul 17, 2025

• [BUG] Agent.__init__() got an unexpected keyword argument 'max_parallel_tools'

  #462 closed Jul 17, 2025

• [BUG] Concurrent Agent Executions Fail with LiteLLM Provider - [Errno 9] Bad file descriptor

  #202 closed Jul 17, 2025

• [BUG] Models - Mistral - Multiple Tool Calls

  #374 closed Jul 17, 2025

• [BUG] LiteLLMModel fails on structured output and uses blocking async calls

  #402 closed Jul 17, 2025

• [BUG] AttributeError: 'NoneType' object has no attribute 'prompt_tokens' with OpenAIModel

  #308 closed Jul 17, 2025

• [FEATURE] Add a data retriever kind-of tool

  #274 closed Jul 17, 2025

• [BUG] Listing MCP tools doesn't paginate through results

  #420 closed Jul 16, 2025

• [BUG] Fix Linting issues

  #128 closed Jul 16, 2025

• [FEATURE] Strongly Type the SDK

  #177 closed Jul 16, 2025

• [FEATURE] Configure the number of responses / chuncks when using BEDROCK KB

  #188 closed Jul 16, 2025

18 Issues opened by 17 people

• [BUG] A2A Doesn't Support Data Types

  #525 opened Jul 23, 2025

• [FEATURE] Support enabling thinking for Sonnet 4/ 3.7 Through LiteLLMModel

  #523 opened Jul 23, 2025

• [FEATURE] Support TypeAdapter-compatible types in `structured_output`

  #522 opened Jul 23, 2025

• [FEATURE] Explicit Model Default Selection for Production Safety

  #521 opened Jul 23, 2025

• [FEATURE] use_aws - Expose stopReason to the client

  #520 opened Jul 23, 2025

• [BUG] use_aws with Nova model

  #519 opened Jul 23, 2025

• [BUG] The content field in the Message object at messages.XX is empty

  #514 opened Jul 22, 2025

• [FEATURE] Support for Type Annotations with Descriptions in @tool Decorator Functions

  #511 opened Jul 21, 2025

• [FEATURE] Applying ConversationManager during the Agent Loop Execution (on specific Hook Events)

  #509 opened Jul 21, 2025

• [FEATURE] Add a parameter for text output instead of using callback_handler behavior

  #506 opened Jul 21, 2025

• [FEATURE] Ability to cancel agent futures

  #501 opened Jul 20, 2025

• [FEATURE] Session Persistence for Graph Agents

  #500 opened Jul 19, 2025

• [FEATURE] Reference-only agent state

  #499 opened Jul 18, 2025

• [FEATURE] Add VPC Endpoint Support to BedrockModel Class

  #496 opened Jul 18, 2025

• [BUG] Regression: ValidationException due to orphaned tool_use blocks after removing message processor

  #495 opened Jul 18, 2025

• [FEATURE] Add support for Amazon Bedrock API Keys authentication for Amazon Bedrock

  #490 opened Jul 17, 2025

• [BUG] TypeError: can only concatenate str (not "int") to str on Bedrock AgentCore

  #488 opened Jul 17, 2025

• [FEATURE] Amazon Bedrock AgentCore Memory Integration for Strands Agents Session Management

  #486 opened Jul 17, 2025

64 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• refactor: update tool registry interface

  #468 commented on Jul 21, 2025 • 21 new comments

• feat(bedrock): support uploads directly from S3

  #310 commented on Jul 23, 2025 • 1 new comment

• [FEATURE] Add support for Bedrock/Anthropic ToolChoice to Structured output

  #256 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support non-stream in OpenAIModel

  #243 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support non-streaming mode for OpenAI model

  #245 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Async Iterator - Typed Events

  #242 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] BedrockSessionSaver

  #261 commented on Jul 22, 2025 • 0 new comments

• [BUG] LiteLLM+Tools not working

  #312 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Add reset() method to Agent class for clearing all agent state between runs

  #329 commented on Jul 22, 2025 • 0 new comments

• [BUG] opt/python/opentelemetry on lambda

  #195 commented on Jul 22, 2025 • 0 new comments

• [BUG] LiteLLM + disable streaming breaks for Azure OpenAI

  #477 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Retries for ServiceUnavailableException (503 errors)

  #370 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Amazon Nova tool caching with cachePoint in messages

  #449 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] MCP 2025-06-18 specification support

  #317 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] AgentRegistry for managing agents

  #390 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Path Prefix Support for A2A Server

  #382 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Configurable HTTP URL for A2A Server

  #383 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Writer model provider

  #227 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Strands MCP client should support prompts/resources/sampling

  #151 commented on Jul 22, 2025 • 0 new comments

• [BUG] openAI stream() method does not enforce "stream=True" in the upstream API call

  #210 commented on Jul 22, 2025 • 0 new comments

• [BUG] OpenAIModel with custom base URL fails when using tool messages

  #215 commented on Jul 22, 2025 • 0 new comments

• [BUG] Protocol Error from Bedrock not handled well

  #114 commented on Jul 23, 2025 • 0 new comments

• feature/ introduce new model - gemini

  #31 commented on Jul 22, 2025 • 0 new comments

• feat: Add list_prompts, get_prompt methods

  #160 commented on Jul 22, 2025 • 0 new comments

• Support for Amazon SageMaker AI endpoints as Model Provider

  #176 commented on Jul 22, 2025 • 0 new comments

• feat: Adding portkey.ai gateway as a custom model

  #197 commented on Jul 22, 2025 • 0 new comments

• feat: Add Twelve Labs Marengo Integration

  #282 commented on Jul 22, 2025 • 0 new comments

• test: adding tests for GitHub official remote server

  #300 commented on Jul 22, 2025 • 0 new comments

• models - llamaapi - async

  #372 commented on Jul 22, 2025 • 0 new comments

• feat(agent): add AgentRegistry for centralized agent management

  #381 commented on Jul 23, 2025 • 0 new comments

• Add baseten integration

  #389 commented on Jul 22, 2025 • 0 new comments

• models: Add support for DeepSeek models

  #428 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Allow setting tool_choice when invoking the agent

  #453 commented on Jul 16, 2025 • 0 new comments

• [BUG] Loss of Temporary System Prompt Override Functionality

  #344 commented on Jul 17, 2025 • 0 new comments

• [FEATURE] Pass Agent objects directly to other Agent objects

  #84 commented on Jul 21, 2025 • 0 new comments

• [FEATURE] Add support of CancellationToken

  #81 commented on Jul 21, 2025 • 0 new comments

• [FEATURE] Enhance ToolRegistry to automatically flatten nested tool collections

  #50 commented on Jul 21, 2025 • 0 new comments

• [FEATURE] LlamaFirewall support in Strands SDK

  #196 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support for multiple MCP servers (and loading from config file)

  #482 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support filtering tools from one or multiple MCP servers

  #481 commented on Jul 22, 2025 • 0 new comments

• [BUG] Graph nodes execute sequentially instead of parallel

  #478 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Allow applications to control thread pools for Agent tool calls.

  #476 commented on Jul 22, 2025 • 0 new comments

• [Refactor] Remove run_async from a2a_client tool

  #471 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Agent as a tool persistence

  #444 commented on Jul 22, 2025 • 0 new comments

• [BUG] Thread conflict error when running multiple agents concurrently

  #333 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Async Initialization

  #330 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Add suite of integration tests across all providers

  #313 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] In-envent-loop cycle context management

  #298 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Allow customer to use custom agent name besides "Strands Agents"

  #277 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Session Persistence

  #246 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Multi-Agent Primitives

  #214 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Implement general-purpose hook/callback system

  #231 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support Human-in-the-Loop (HIL) capabilities

  #204 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Improved MCP DX to prevent common context manager scoping issue

  #198 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Add configurable limits for cycles, token counts, and overall runtime

  #191 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Implement native async iterator support

  #83 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Add Database Session Management

  #33 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Pause and resume a workflow

  #36 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Agent to Agent Communication (A2A)

  #24 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Community-driven session management

  #57 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Add cross-platform uvx command support to StdioServerParameters

  #92 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Support for async Tool

  #116 commented on Jul 22, 2025 • 0 new comments

• [EPIC] Implement native async iterator support

  #129 commented on Jul 22, 2025 • 0 new comments

• [FEATURE] Allow tool consent in async frameworks

  #457 commented on Jul 22, 2025 • 0 new comments