Package index • llmr

Core Components

Essential functions for creating and managing agents and workflows

new_agent(): Create a new agent

new_workflow(): Create a new workflow

new_provider() new_anthropic() new_openai(): Create a provider

new_message(): Create a message

add_tool(): Add a tool to an agent

Request & Response Handling

Functions for making requests and handling responses

request(): Make a request to an LLM provider

Workflow Management

Functions for building and executing workflows

execute(): Execute a workflow

step(): Create a workflow step

when(): Create conditional branching in workflows

`%->%`: Workflow pipe operator

Message Management

Functions for working with messages

append_message(): Add a message to the list

clear_messages(): Clear messages

get_messages(): Get messages

get_last_message(): Get last message

Provider Configuration

Functions for configuring LLM providers

set_api_key(): Set the key for an LLM provider

set_max_tokens(): Set the token for an LLM provider

set_max_tool_calls(): Set the maximum number of tool calls for an LLM provider

set_model(): Set the model for an LLM provider

set_retry(): Set Retry Parameters for LLM Provider

set_system_prompt(): Set the system prompt for an LLM provider

set_temperature(): Set the temperature for an LLM provider

set_version(): Set the version for an LLM provider

MCP Integration

Functions for Memory Client Provider integration

register_mcp(): Register a memory client provider

Human-in-the-Loop Approval

Functions for requiring human approval before tool execution

set_approval_callback(): Set approval callback for human-in-the-loop tool execution

prompt_human_approval(): Interactive human approval prompt for tool execution

batch_approval_interface(): Batch approval interface for multiple tool calls

smart_approval_callback(): Smart approval callback that uses stored preferences

clear_tool_approvals(): Clear all stored tool approval preferences

list_tool_approvals(): List current tool approval preferences