Set up messages for the agent execution.

Whether this kickoff should defer final flow trace finalization.

Whether to run PlannerObserver LLM calls after each step. Explicit ``observe_steps=False`` disables observation at any effort level.

Check if the model supports native structured outputs. Native structured outputs are only available for Claude 4.5 models (Sonnet 4.5, Opus 4

Build an in-memory tar archive via `build(tf)` and return it for reading.

Parse streamed observation JSON and update the exact step it names.

Remove key tracking for a provider.

Convert a UUID string to a stable Qdrant point ID. Falls back to hashing for non-UUID strings.

Validate keyword arguments against args_schema if present. Args: kwargs: The keyword arguments to validate. Returns:

Call the LLM with the given messages. Args: messages: Input messages for the LLM. Can be a string or list of

Get merged crew and task files asynchronously. Task files override crew files with the same name. Args: crew_id: Unique identifier f

Get all cached uploads for a provider. Args: provider: The provider name. Returns: List of cached uploads fo

Query across all knowledge sources asynchronously. Args: query: List of query strings. results_limit: Maximum number

Override streaming for ``llm`` within the current call scope.

Clear all registered global before_tool_call hooks. Returns: Number of hooks that were cleared Example: >>> register_before_

Remove the stored token file (no-op if absent).

Clear all user settings.

Nearest candidate name to a likely typo, if one is close enough.

Start the MCP server process and establish connection. Returns: Self for method chaining. Raises: Connection

Printable text content for chunk-like consumers.

Convert oneOf to anyOf for OpenAI compatibility. OpenAI's Structured Outputs support anyOf better than oneOf. This recursively converts all o

Convert a result string to a Pydantic model or JSON using instructions. Args: result: The result string to convert. model: The Py

Create a client from configuration using the appropriate factory. Args: config: The RAG client configuration. Returns: The c

Create state for a scoped public frame stream.

Create a mock LLM that returns a simple response. The mock includes all attributes required by the telemetry system, particularly the 'model'

Create research tools for testing structured planning.

Create a configured FileResolver. Args: provider: Optional provider name to load default threshold from constraints. prefer_uploa

Ensure all properties are in the required array for OpenAI strict mode. OpenAI's strict structured outputs require all properties to be listed

Records the coding tool environment context.

Fetch latest updates from the remote.

Return ``stem.jsonc`` or ``stem.json``, preferring JSONC.

Create a MergeAgentHandlerTool from a tool name. Args: tool_name: Name of the tool (e.g., "linear__create_issue")

Generate the ChatInputs required for the crew by analyzing the tasks and agents. Args: crew: The crew object containing tasks and agents.

Return the saved authentication token; raise ``AuthError`` if missing.

Return the cached skill directory path if it exists, else None.

Read message history from flow state or the internal fallback buffer.

Get all failed todos. Returns: List of failed TodoItem objects.

Return the project name from ``pyproject.toml``.

Return a single record by ID, or None if not found. Args: record_id: The unique ID of the record. Returns: T

Get information about a scope. Args: scope: The scope path. Returns: ScopeInfo with record count, categories

Get supported MIME type prefixes for a provider. Args: provider: Provider name string. api: Optional API variant (e.g., "response

Handle OutputParserError by updating messages and formatted_answer. Args: e: The OutputParserError that occurred messages: List o

Check if there are any built-in tool outputs.

Check if this is first time and initialize collection.

Execute the agent with given inputs. Args: inputs: Input dictionary containing prompt variables. Returns: Di

Check if code is currently running inside an asyncio event loop. This is used to detect when code is being called from within an async context

Return whether an error means native tool calling is unavailable.

Return True if *value* looks like a registry reference (@org/name).

List available templates with an interactive selector to install.

List all available triggers from integrations.

Load all stored task outputs. Returns: List of task output dictionaries, or None if no outputs exist.

Load a ``Crew`` from an in-memory JSON/YAML crew definition.

Load content from a URL. Args: url: The URL to fetch content from kwargs: Additional keyword arguments (can include 'headers' ove

Promote a skill to INSTRUCTIONS disclosure level. Reads the full SKILL.md body text. Args: skill: Skill at METADATA level. Retu

Load skill inputs into de-duplicated Skill objects. Preserves first-seen order when multiple inputs resolve to the same skill name. Registry

Log successful LLM API call and track token usage.

Mark a todo as failed by step number.

Normalize a pyproject project.name into a Python package directory name. Mirrors the rules in ``crewai.cli.create_crew.create_crew`` so the v

Get the Git repository's remote URL.

Parse agent output text into AgentAction or AgentFinish. Expects output to be in one of two formats. If the output signals that an action sh

Read and parse a SKILL.md file. Args: path: Path to the SKILL.md file. Returns: Tuple of (SkillFrontmatter, body text).

Parse tools to be used for the task. Args: tools: List of tools to parse. Returns: List of structured tools. Raises:

Create interactive HTML visualization of Flow structure. Args: filename: Output HTML filename (default: "crewai_flow.html").

Pop an event entry from the scope stack.

Prepare crew for kickoff execution. Handles before callbacks, event emission, task handler reset, input interpolation, task callbacks, agent

Process the guardrail for the agent output. Args: output: The output to validate with the guardrail guardrail: The guardrail to v

Prompt for each name and return ``{name: value}``. ``describe(name)`` returns an optional hint shown dim above the prompt (used by flows to s

Remove old checkpoints, keeping at most *max_keep* per branch. Args: location: The storage destination passed to ``checkpoint``.

Register a cleanup callback on the streaming output. Ensures the event handler is unregistered even if aclose()/close() is called before iter

Request human input during tool hook execution. This method pauses live console updates, displays a prompt to the user, waits for the

Reset all settings to default values.

Return the Python symbol named by a project definition field.

Sanitize tool name for API compatibility.