1.7.2

What's Changed

Bug Fixes

• Resolve connection issues (#4129)

Documentation

• Update api-reference/status docs page

Contributors

@greysonlalonde, @heitorado, @lorenzejay, @lucasgomide

1.7.1

What's Changed

Improvements

• Add --no-commit flag to bump command
• Use JSON schema for tool argument serialization

Bug Fixes

• Fix error message display from response when tool repository login fails
• Fix graceful termination of future when executing a task asynchronously
• Fix task ordering by adding index
• Fix platform compatibility checks for Windows signals
• Fix RPM controller timer to prevent process hang
• Fix token usage recording and validate response model on stream

Documentation

• Add translated documentation for async
• Add documentation for AOP Deploy API
• Add documentation for the agent handler connector
• Add documentation on native async

Contributors

@Llamrei, @dragosmc, @gilfeig, @greysonlalonde, @heitorado, @lorenzejay, @mattatcha, @vinibrsl

1.7.0

What's Changed

Features

• Add async flow kickoff
• Add async crew support
• Add async task support
• Add async knowledge support
• Add async memory support
• Add async support for tools and agent executor; improve typing and docs
• Implement a2a extensions API and async agent card caching; fix task propagation & streaming
• Add native async tool support
• Add async llm support
• Create sys event types and handler

Bug Fixes

• Fix issue to ensure nonetypes are not passed to otel
• Fix deadlock in token store file operations
• Fix to ensure otel span is closed
• Use HuggingFaceEmbeddingFunction for embeddings, update keys and add tests
• Fix to ensure supports_tools is true for all supported anthropic models
• Ensure hooks work with lite agents flows

Contributors

@greysonlalonde, @lorenzejay

1.6.1

What's Changed

Bug Fixes

• Fix ChatCompletionsClient call to ensure proper functionality (#3910)
• Ensure async methods are executable for annotations
• Fix parameters in RagTool.add, add typing, and tests (#3979)
• Remove invalid parameter from SSE client (#3980)
• Erase 'oauth2_extra' setting on 'crewai config reset' command

Refactoring

• Enhance model validation and provider inference in LLM class (#3976)

Contributors

@Vidit-Ostwal, @greysonlalonde, @heitorado, @lorenzejay

1.6.0

What's Changed

Features

• Add streaming result support to flows and crews
• Add gemini-3-pro-preview
• Support CLI login with Entra ID
• Add Merge Agent Handler tool
• Enhance flow event state management

Bug Fixes

• Ensure custom rag store persist path is set if passed
• Ensure fuzzy returns are more strict and show type warning
• Re-add openai response_format parameter and add test
• Fix rag tool embeddings configuration
• Ensure flow execution start panel is not shown on plot

Documentation

• Update references from AMP to AOP in documentation
• Update AMP to AOP

Contributors

@Vidit-Ostwal, @gilfeig, @greysonlalonde, @heitorado, @joaomdmoura, @lorenzejay, @markmcd

0.203.2

What's Changed

• chore: update hotfix crewAI version 0.203.1 -> 0.203.2 by @lorenzejay in #3963

Full Changelog: 0.203.1...0.203.2

1.5.0

What's Changed

Features

• Add a2a trust remote completion status flag
• Fetch and store more data about Okta authorization server
• Implement before and after LLM call hooks in CrewAgentExecutor
• Expose messages to TaskOutput and LiteAgentOutputs
• Enhance schema description of QdrantVectorSearchTool

Bug Fixes

• Ensure tracing instrumentation flags are correctly applied
• Fix custom tool documentation links and add Mintlify broken links action

Documentation

• Enhance task guardrail documentation with LLM-based validation support

Contributors

@danielfsbarreto, @greysonlalonde, @heitorado, @lorenzejay, @theCyberTech

1.4.1

What's Changed

Bug Fixes

• Fix handling of agent max iterations
• Resolve routing issues for LLM model syntax to respected providers

@greysonlalonde

1.4.0

What's Changed

Features

• Add support for non-AST plot routes
• Implement first-class support for MCP
• Add Pydantic validation dunder to BaseInterceptor
• Add support for LLM message interceptor hooks
• Cache i18n prompts for efficient use
• Enhance QdrantVectorSearchTool

Bug Fixes

• Fix issues with keeping stopwords updated
• Resolve unpickleable values in flow state
• Ensure lite agents course-correct on validation errors
• Fix callback argument hashing to ensure caching works
• Allow adding RAG source content from valid URLs
• Make plot node selection smoother
• Fix duplicating document IDs for knowledge

Refactoring

• Improve MCP tool execution handling with concurrent futures
• Simplify flow handling, typing, and logging; update UI and tests
• Refactor stop word management to a property

Documentation

• Migrate embedder to embedding_model and require vectordb across tool docs; add provider examples (en/ko/pt-BR)

Contributors

@danielfsbarreto, @greysonlalonde, @lorenzejay, @lucasgomide, @tonykipkemboi

1.3.0

What's Changed

Features

• Refactor flow handling, typing, and logging
• Enhance QdrantVectorSearchTool

Bug Fixes

• Fix Firecrawl tools and add tests
• Refactor use_stop_words to property and add check for stop words

Documentation

• Migrate embedder to embedding_model and require vectordb across tool docs
• Add provider examples in English, Korean, and Portuguese

Refactoring

• Improve flow handling and UI updates

Contributors

@danielfsbarreto, @greysonlalonde, @lorenzejay, @lucasgomide, @tonykipkemboi