CrewAI
AI & LLMsFramework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
Release History
View all versions →1.15.1811 fixes9 featuresThis release promotes conversational flows to stable, enhances their documentation and APIs, and introduces several bug fixes including improved tool result handling and correct recording of task failures. It also refines project ID handling and LLM configuration for conversational declarations.
1.15.175 fixes6 featuresThis release introduces declarative conversational flows, enhancing how conversational modes are driven and managed. It also includes several bug fixes related to server names, agent scopes, tool error attribution, SSRF checks, and native tool calls.
1.15.163 fixes4 featuresThis release introduces enhanced execution context management with UUID support and improved error/deployment tracking. It also includes several bug fixes for span recording, MySQL validation, and turn status.
1.15.154 fixes1 featureThis release introduces new reporting features for flow outcomes and signals, alongside several bug fixes including security vulnerability patches for torch and gitpython. It also standardizes CLI flags to kebab-case.
1.15.141 featureThis release introduces a split of runtime context from the coding agent and adds project ID functionality. Documentation has also been updated.
1.15.134 fixesThis release includes several bug fixes, including improvements to LiteLLM-routed models, LLM event-bus mocks, and Anthropic cache token usage. It also addresses a security vulnerability by bumping the h2 dependency.
1.15.121 fix4 featuresThis release introduces new features like the URLReadTool and unifies scaffolding under a new command. It also includes bug fixes for name collision errors and documentation updates.
1.15.114 fixes4 featuresThis release introduces new features for telemetry tracking, enterprise account linking, and an IBM Db2 search tool. It also includes several bug fixes related to security advisories, error reporting, and CI skipping.
1.15.101 featureThis release introduces the ability to collect skill usage events. Documentation has also been updated, including the removal of migrated AMP documentation and updated security reporting guidelines.
1.15.93 featuresThis release introduces new features for handling tool failures, emitting flow failure events, and implementing progressive disclosure for skills. It also includes documentation updates.
1.15.82 fixes1 featureThis release introduces the WaitTool for pausing long-running jobs and fixes issues with the FileWriterTool. It also clarifies the requirement for the E2B_API_KEY environment variable.
1.15.75 fixes1 featureThis release includes several bug fixes, including improvements to tool calling and model routing. It also introduces new observability features by emitting skill usage events at runtime.
1.15.7a15 fixes1 featureThis release includes several bug fixes, including improvements to skill resolution, error recovery, and API routing. It also introduces enhanced observability by emitting skill usage events at runtime.
1.15.65 fixesThis release includes several bug fixes, including improvements to Anthropic tool-use block detection, preservation of tool schema property names, and better handling of crew/flow execution failures. It also addresses issues with async agent retrieval and dependency resolution.
1.15.51 featureThis release introduces a new feature for authenticating skill registry downloads. It also includes documentation updates for version v1.15.4.
1.15.41 featureThe Skills Repository has been promoted out of experimental status. Documentation has also been updated to include information on Flows in Studio.
1.15.39 fixes5 featuresThis release introduces new features for interception points and TUI execution, alongside several bug fixes to improve stability and functionality. Tool result caching is now opt-in by default.
1.15.3a21 fixThis release includes a bug fix for event synchronization and updates the setuptools dependency to address a security vulnerability.
1.15.3a19 fixes6 featuresThis release introduces new features for API clients and execution hooks, including organization ID parameter, step interception points, and TUI support for declarative flows. It also addresses several bugs related to agent attributes, tool execution, and history handling.
1.15.27 fixes11 featuresThis release introduces significant new features for flow management, including dynamic model pulling and inline skill definitions. It also addresses several bug fixes related to caching, dependency vulnerabilities, and flow execution.
1.15.2a21 fix5 featuresThis release introduces new features related to flow agents and bedrock extras, alongside a bug fix rejecting self-listening flow methods.
1.15.2a11 fix5 featuresThis release introduces new features like inline skill definitions and flow protocol support, alongside a fix for documentation navigation issues.
1.15.14 fixes3 featuresThis release introduces new features like Git initialization for generated projects and explicit CrewAI project definitions. Several bugs were fixed, including issues with deployment links, JSON template rendering, and SSRF vulnerabilities.
1.15.1a13 fixes3 featuresThis release introduces new telemetry tracking, enforces explicit project definitions, and fixes several bugs related to JSON rendering and security vulnerabilities.
1.15.07 fixes10 featuresThis release introduces significant features for conversational flows, including CLI support and declarative loading, alongside various bug fixes and internal refactoring for better flow state management.
Common Errors
ModuleNotFoundError3 reportsModuleNotFoundError in CrewAI usually indicates a missing or incorrectly installed dependency. To fix this, ensure you've installed all required packages using `pip install crewai` (or `pip install crewai[tools]` if you need the tools). For specific missing modules like `litellm`, install them directly using `pip install litellm`.
FileNotFoundError3 reportsFileNotFoundError in crewai usually arises from incorrect file paths, especially when the program cannot locate a specified file or directory, often due to relative paths or operating system differences. Fix this by using absolute paths where possible, validating file existence with `os.path.exists()` before access, and normalizing paths using `os.path.normpath()` to handle OS differences, especially on Windows with long paths by using the `\\?\` prefix.
OutputParserException2 reportsOutputParserException in CrewAI arises when the LLM's output doesn't conform to the expected format, causing parsing failures. Correct this by ensuring the `format_instructions` in your prompt are very clear and align with your OutputParser's requirements. Furthermore, implement robust error handling within the parsing logic to gracefully manage unexpected formats, potentially returning a default or partially parsed response.
NotImplementedError2 reportsBadRequestError2 reportsBadRequestError in crewAI often arises from invalid input sent to the LLM provider, such as empty content or malformed prompts. To fix this, ensure prompt strings are properly formatted and non-empty before sending them to the LLM, and strip any trailing whitespace. Validation and sanitization of input strings will prevent such errors.
RateLimitError1 reportRateLimitError in crewAI usually occurs when the language model API is called too frequently, exceeding the allowed requests per time unit. Implement exponential backoff with retry logic around API calls to pause and retry after rate limits are hit. Consider using a rate limiting library or adjusting the agent's task execution frequency to stay within the API's usage limits.
Related AI & LLMs Packages
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.
🦜🔗 The platform for reliable agents.
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
LLM inference in C/C++
GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
Subscribe to Updates
Get notified when new versions are released