Skip to contents

Chat with CassidyAI

Source: R/chat-core.R
cassidy_chat.Rd

Send a message to a CassidyAI assistant and get a response. This is the main function for interacting with CassidyAI in a conversational way.

Usage

cassidy_chat(
  message,
  assistant_id = Sys.getenv("CASSIDY_ASSISTANT_ID"),
  thread_id = NULL,
  context = NULL,
  api_key = Sys.getenv("CASSIDY_API_KEY"),
  timeout = 120
)

Arguments

message

Character. The message to send to the assistant.

assistant_id

Character. The CassidyAI assistant ID. Defaults to the CASSIDY_ASSISTANT_ID environment variable.

thread_id

Character or NULL. An existing thread ID to continue a conversation. If NULL (default), a new thread is created.

context

Optional context object from cassidy_context_project(), cassidy_describe_df(), or a custom context object with a text element. Context is sent once at thread creation for efficiency.

api_key

Character. Your CassidyAI API key. Defaults to the CASSIDY_API_KEY environment variable.

timeout

Numeric. Request timeout in seconds. Default is 120.

Value

A cassidy_chat S3 object containing:

thread_id

The thread ID (save this to continue the conversation)

response

The assistant's response (a cassidy_response object)

message

Your original message

Details

If no thread_id is provided, a new conversation thread is created automatically. To continue an existing conversation, pass the thread_id from a previous call.

See also

Other chat-functions: cassidy_continue(), cassidy_session(), cassidy_write_code(), cassidy_write_file(), chat(), chat_text()

Examples

if (FALSE) { # \dontrun{
  # Simple one-off question
  result <- cassidy_chat("What is the tidyverse?")
  print(result)

  # With project context
  ctx <- cassidy_context_project()
  result <- cassidy_chat("Help me understand this project", context = ctx)

  # With data frame context
  desc <- cassidy_describe_df(mtcars)
  result <- cassidy_chat("What analyses would you recommend?", context = desc)

  # Continue the conversation (context already set)
  result2 <- cassidy_continue(result, "Tell me more")
} # }