Represents an agent in a system. Each agent has a role, a goal, a backstory, and an optional language model (llm). The agent can also have me

Class that represents a task to be executed. Each task must have a description, an expected output and an agent responsible for execution. A

Represents a group of agents, defining how they should collaborate and the tasks they should perform. Attributes: tasks: list of

File created from raw bytes content.

File representing an image. Supports common image formats: PNG, JPEG, GIF, WebP, BMP, TIFF, SVG.

OAuth2 provider configuration.

Unified memory: standalone, LLM-analyzed, with intelligent recall flow. Works without agent/crew. Uses LLM to infer scope, categories, importance

Class that represents the result of a task. Attributes: description: Description of the task name: Optional name of the task

Displays text content.

OpenAI native completion implementation. This class provides direct integration with the OpenAI Python SDK, supporting both Chat Completions

Represents the final answer from an agent.

SQLite-based implementation of flow state persistence. This class provides a simple, file-based persistence implementation using SQLite. It's

CrewAI platform settings persisted to ``~/.config/crewai/settings.json``.

A single todo item representing a step in the execution plan.

File representing a text document. Supports common text formats: TXT, MD, RST, CSV, JSON, XML, YAML, HTML.

Okta OAuth2 provider implementation.

Context capturing everything needed to resume a paused flow. When a flow is paused waiting for async human feedback, this dataclass stores al

Configuration for automatic checkpointing. When set on a Crew, Flow, or Agent, checkpoints are written automatically whenever the specified e

Context object passed to tool call hooks. Provides hooks with access to the tool being called, its input, the agent/task/crew context, and th

Collection of todos for tracking plan execution.

Manages encrypted token storage on disk under platform-appropriate paths.

File representing a PDF document.

File referenced by URL. For providers that support URL references, the URL is passed directly. For providers that don't, content is fetched o

Configuration for agent planning/reasoning before task execution. This allows users to customize the planning behavior including prompts, ite

Class that represents the usage of a tool by an agent. Attributes: task: Task being executed. tools_handler: Tools handler that

Generic file that auto-detects the appropriate type. Use this when you don't want to specify the exact file type. The content type is automat

Context object passed to LLM call hooks. Provides hooks with complete access to the execution state, allowing modification of messages, respo

A knowledge source that stores and queries plain text content using embeddings.

Trace collection listener that orchestrates trace collection.

Track usage metrics for crew execution. Attributes: total_tokens: Total number of tokens used. prompt_tokens: Number of tokens us

Public Flow class with experimental conversational extension behavior.

Extends Storage to handle embeddings for memory entries, improving search efficiency.

Persists runtime state checkpoints as JSON files on the local filesystem.

A single memory entry stored in the memory system.

Azure AI Inference native completion implementation. This class provides direct integration with the Azure AI Inference Python SDK, offering

Runs the full pre-deploy validation suite against a project directory.

A tool for performing semantic searches on SingleStore database tables. This tool provides a safe interface for executing SELECT and SHOW queries

YAML frontmatter from a SKILL.md file. Attributes: name: Unique skill identifier (1-64 chars, lowercase alphanumeric + hyphens).

Configuration for exposing a Crew or Agent as an A2A server. All fields correspond to A2A AgentCard fields. Fields like name, description, an

User-facing error raised while loading JSON-first crew projects.

Configuration for CrewAI security features. This class manages security settings for CrewAI agents, including: - Authentication credenti

Class that converts text into either pydantic or json.

Resolves files to their delivery format based on provider capabilities. Decides whether to use inline base64, raw bytes, or file upload based on:

Entra ID (Azure AD) OAuth2 provider implementation.

Exception raised when the context length of a language model is exceeded. Attributes: original_error_message: The original error message

Event emitted when a tool execution is started

Streaming output wrapper for crew execution. Provides both sync and async iteration over stream chunks, with access to the final CrewOutput v

Knowledge is a collection of sources and setup for the vector store to save and query relevant context. Args: sources: list[BaseKnowl

Event emitted when a streaming chunk is received

Snowflake Cortex REST API native completion implementation. Snowflake exposes an OpenAI-compatible Chat Completions endpoint at ``/api/v2/cor

Handle anonymous telemetry for the CrewAI package. Attributes: ready: Whether telemetry is initialized and ready. trace_set: Whet

Single responsibility: Manage batches and event buffering

Processes files according to provider constraints and per-file mode mode. Validates files against provider-specific limits and optionally transfo

Keycloak OAuth2 provider implementation.

File representing a video file. Supports common video formats: MP4, AVI, MKV, MOV, WebM, FLV, WMV.

Mock input provider that returns pre-configured responses.

Qdrant implementation of the BaseClient protocol. Provides vector database operations for Qdrant, supporting both synchronous and asynchronou

Async cache for tracking uploaded files using aiocache. Supports in-memory caching by default, with optional Redis backend for distributed se

WorkOS OAuth2 provider implementation.

File representing an audio file. Supports common audio formats: MP3, WAV, OGG, FLAC, AAC, M4A, WMA.

Directed record of execution events with O(1) node lookup. Events are added via :meth:`add` which automatically wires edges based on the even

Signal that flow execution should pause for async human feedback. When raised by a provider, the flow framework will: 1. Stop execution at th

Class that wraps an agent LLM with instructor for structured output generation. Attributes: content: The content to be processed

OpenAI-compatible completion implementation. This class provides support for various OpenAI-compatible APIs by automatically configuring the

A loaded Agent Skill with progressive disclosure support. Attributes: frontmatter: Parsed YAML frontmatter. instructions: Full SK

A task that can be conditionally executed based on the output of another task. This task type allows for dynamic workflow execution based on the

Enhanced embedding service that uses CrewAI's existing embedding providers. Supports multiple providers: - openai: OpenAI embeddings (te

A class for generating and managing unique identifiers for agents. Each agent has dual identifiers: - Human-readable ID: For debugging and re

Constraints for image files. Attributes: max_size_bytes: Maximum file size in bytes. max_width: Maximum image width in pixels.

Aggregates validation errors found without executing a JSON project.

LanceDB-backed storage for the unified memory system.

Manages the lifecycle of an MCP server and make its tools available to CrewAI. Note: tools can only be accessed after the server has been started

A single step in the reasoning plan.

Complete set of constraints for a provider. Attributes: name: Provider name identifier. image: Image file constraints. pd

Base streaming chunk with full context. Attributes: content: The streaming content (text or partial content) chunk_type: Type of

Auth0 OAuth2 provider implementation.

Raised when a method is called with the wrong client type. Typically used when a sync method is called with an async client, or vice versa.

Builds platform tools from remote action schemas.

Event emitted when a memory save operation is started

Client for working with the CrewAI+ API.

Manages the global skill cache at ~/.crewai/skills/.

Constraints for video files. Attributes: max_size_bytes: Maximum file size in bytes. max_duration_seconds: Maximum video duration

Simple dummy interceptor for testing.

Event emitted when a LLM call completes

Wrapper for Merge Agent Handler tools. This tool allows CrewAI agents to execute tools from Merge Agent Handler, which provides secure a

Read-write lock for managing concurrent read and exclusive write access. Allows multiple threads to acquire read locks simultaneously, but ensure

Event emitted when a tool execution encounters an error

Event emitted when a tool execution is completed