llm_provider_ellmer: Create a new LLM provider from an 'ellmer::chat()' object

In tidyprompt: Prompt Large Language Models and Enhance Their Functionality

Create a new LLM provider from an ellmer::chat() object

Description

This function creates a llm_provider from an ellmer::chat() object. This allows the user to use the various LLM providers which are supported by the 'ellmer' R package, including respective configuration and features.

Please note that this function is experimental. This provider type may show different behavior than other LLM providers, and may not function optimally.

Usage

llm_provider_ellmer(chat, verbose = getOption("tidyprompt.verbose", TRUE))

Arguments

chat	An ellmer::chat() object (e.g., ellmer::chat_openai())
verbose	A logical indicating whether the interaction with the llm_provider should be printed to the console. Default is TRUE

Details

Unlike other LLM provider classes, LLM provider settings need to be managed in the ellmer::chat() object (and not in the ⁠$parameters⁠ list). ⁠$get_chat()⁠ and ⁠$set_chat()⁠ may be used to manipulate the chat object.

A special parameter ⁠$.ellmer_structured_type⁠ may be set in the ⁠$parameters⁠ list; this parameter is used to specify a structured output format. This should be a 'ellmer' structured type (e.g., ellmer::type_object; see https://ellmer.tidyverse.org/articles/structured-data.html). answer_as_json() sets this parameter to obtain structured output (it is not recommended to set this parameter manually, but it is possible).

Value

An llm_provider with api_type = "ellmer"

See Also

Other llm_provider: llm_provider-class, llm_provider_google_gemini(), llm_provider_groq(), llm_provider_mistral(), llm_provider_ollama(), llm_provider_openai(), llm_provider_openrouter(), llm_provider_xai()

Examples

# Various providers:
ollama <- llm_provider_ollama()
openai <- llm_provider_openai()
openrouter <- llm_provider_openrouter()
mistral <- llm_provider_mistral()
groq <- llm_provider_groq()
xai <- llm_provider_xai()
gemini <- llm_provider_google_gemini()

# From an `ellmer::chat()` (e.g., `ellmer::chat_openai()`, ...):
## Not run: 
ellmer <- llm_provider_ellmer(ellmer::chat_openai())

## End(Not run)

# Initialize with settings:
ollama <- llm_provider_ollama(
  parameters = list(
    model = "llama3.2:3b",
    stream = TRUE
  ),
  verbose = TRUE,
  url = "http://localhost:11434/api/chat"
)

# Change settings:
ollama$verbose <- FALSE
ollama$parameters$stream <- FALSE
ollama$parameters$model <- "llama3.1:8b"

## Not run: 
# Try a simple chat message with '$complete_chat()':
response <- ollama$complete_chat("Hi!")
response
# $role
# [1] "assistant"
#
# $content
# [1] "How's it going? Is there something I can help you with or would you like
# to chat?"
#
# $http
# Response [http://localhost:11434/api/chat]
# Date: 2024-11-18 14:21
# Status: 200
# Content-Type: application/json; charset=utf-8
# Size: 375 B

# Use with send_prompt():
"Hi" |>
  send_prompt(ollama)
# [1] "How's your day going so far? Is there something I can help you with or
# would you like to chat?"

## End(Not run)

tidyprompt documentation built on Aug. 25, 2025, 5:11 p.m.