Nothing
R/chatLLM.R
In chatLLM: A Flexible Interface for 'LLM' API Interactions
Defines functions
call_llm
list_models
get_all_github_models
get_dashscope_models
get_deepseek_models
get_anthropic_models
get_groq_models
get_gemini_models
get_grok_models
get_openai_models
parse_response
get_api_key
get_default_model
Documented in
call_llm
list_models
###############################################################################
# chatLLM helpers + call_llm() #
# Requires: httr (>= 1.4.6), jsonlite #
###############################################################################
# ------------------------------------------------------------------------------
#' @importFrom httr POST add_headers http_error content
#' @importFrom jsonlite toJSON fromJSON
#' @importFrom httr GET timeout status_code
#' @importFrom stats setNames
library(httr)
library(jsonlite)
###############################################################################
# 1. Provider defaults #
###############################################################################
get_default_model <- function(provider) {
switch(
tolower(provider),
"openai" = "gpt-3.5-turbo",
"groq" = "meta-llama/llama-4-scout-17b-16e-instruct",
"anthropic" = "claude-3-7-sonnet-20250219",
"deepseek" = "deepseek-chat",
"dashscope" = "qwen-plus-latest",
"github" = "openai/gpt-4.1",
"gemini" = "gemini-2.0-flash",
"grok" = "grok-3-latest",
stop("No default model for provider: ", provider)
)
}
###############################################################################
# 2. API-key helper #
###############################################################################
get_api_key <- function(provider, api_key = NULL) {
env_var <- switch(
tolower(provider),
"openai" = "OPENAI_API_KEY",
"groq" = "GROQ_API_KEY",
"anthropic" = "ANTHROPIC_API_KEY",
"deepseek" = "DEEPSEEK_API_KEY",
"dashscope" = "DASHSCOPE_API_KEY",
"github" = "GH_MODELS_TOKEN",
"gemini" = "GEMINI_API_KEY",
"grok" = "XAI_API_KEY",
stop("Unknown provider: ", provider)
)
if (is.null(api_key)) api_key <- Sys.getenv(env_var)
if (!nzchar(api_key))
stop(sprintf("API key not found for %s. Set %s or pass `api_key`.",
provider, env_var))
api_key
}
###############################################################################
# 3. Parse chat-completion responses #
###############################################################################
parse_response <- function(provider, parsed) {
switch(
tolower(provider),
"openai" = parsed$choices[[1]]$message$content,
"groq" = parsed$choices[[1]]$message$content,
"anthropic" = parsed$content[[1]]$text,
"deepseek" = parsed$choices[[1]]$message$content,
"dashscope" = parsed$choices[[1]]$message$content,
"github" = parsed$choices[[1]]$message$content,
"gemini" = parsed$choices[[1]]$message$content,
"grok" = parsed$choices[[1]]$message$content,
stop("Parsing not implemented for provider: ", provider)
)
}
###############################################################################
# 4. Model-catalog helpers (one per provider) #
###############################################################################
get_openai_models <- function(token = Sys.getenv("OPENAI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.openai.com/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_grok_models <- function(token = Sys.getenv("XAI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.x.ai/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_gemini_models <- function(token = Sys.getenv("GEMINI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://generativelanguage.googleapis.com/v1beta/openai/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_groq_models <- function(token = Sys.getenv("GROQ_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.groq.com/openai/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_anthropic_models <- function(token = Sys.getenv("ANTHROPIC_API_KEY"),
anthropic_api_version = "2023-06-01",
limit = 1000) {
if (!nzchar(token)) return(character())
r <- tryCatch(
GET("https://api.anthropic.com/v1/models",
add_headers(`x-api-key` = token,
`anthropic-version` = anthropic_api_version),
query = list(limit = limit),
timeout(60)),
error = function(e) NULL)
if (is.null(r) || http_error(r)) return(character())
p <- content(r, "parsed")
if (!is.null(p$data))
return(vapply(p$data, `[[`, character(1), "id"))
if (!is.null(p$models))
return(vapply(p$models, `[[`, character(1), "name"))
character()
}
get_deepseek_models <- function(token = Sys.getenv("DEEPSEEK_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.deepseek.com/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_dashscope_models <- function(token = Sys.getenv("DASHSCOPE_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://dashscope-intl.aliyuncs.com/compatible-mode/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_all_github_models <- function(token = Sys.getenv("GH_MODELS_TOKEN"),
api_version = "2022-11-28") {
if (!nzchar(token)) return(character())
r <- GET("https://models.github.ai/catalog/models",
add_headers(Accept = "application/vnd.github+json",
Authorization = paste("Bearer", token),
`X-GitHub-Api-Version` = api_version),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed", simplifyVector = FALSE),
`[[`, character(1), "id")
}
###############################################################################
# 5. list_models() #
###############################################################################
#' List Available Models for Supported Providers
#'
#' @name list_models
#'
#' @description
#' Retrieve the catalog of available model IDs for one or all supported
#' chat - completion providers. Useful for discovering active models and
#' avoiding typos or deprecated defaults.
#'
#' Supported providers:
#' \itemize{
#' \item \code{"openai"} - OpenAI Chat Completions API
#' \item \code{"groq"} - Groq OpenAI - compatible endpoint
#' \item \code{"anthropic"} - Anthropic Claude API
#' \item \code{"deepseek"} - DeepSeek chat API
#' \item \code{"dashscope"} - Alibaba DashScope compatible API
#' \item \code{"github"} - GitHub Models OpenAI - compatible API
#' \item \code{"all"} - Fetch catalogs for all of the above
#' }
#'
#' @param provider Character. One of \code{"github"}, \code{"openai"},
#' \code{"groq"}, \code{"anthropic"}, \code{"deepseek"},
#' \code{"dashscope"} or \code{"all"}. Case - insensitive.
#' @param ... Additional arguments passed to the per - provider helper
#' (e.g. \code{limit} for Anthropic, or \code{api_version} for GitHub).
#' @param github_api_version Character. Header value for
#' \code{X - GitHub - Api - Version} (GitHub Models). Default \code{"2022 - 11 - 28"}.
#' @param anthropic_api_version Character. Header value for
#' \code{anthropic - version} (Anthropic). Default \code{"2023 - 06 - 01"}.
#'
#' @return
#' If \code{provider != "all"}, a character vector of model IDs for that
#' single provider. If \code{provider == "all"}, a named list of character
#' vectors, one per provider.
#'
#' @examples
#' \dontrun{
#' Sys.setenv(OPENAI_API_KEY = "sk-...")
#' openai_models <- list_models("openai")
#' head(openai_models)
#'
#' Sys.setenv(ANTHROPIC_API_KEY = "sk-...")
#' anthro_models <- list_models("anthropic", anthropic_api_version = "2023-06-01")
#'
#' Sys.setenv(GH_MODELS_TOKEN = "ghp-...")
#' github_models <- list_models("github", github_api_version = "2022-11-28")
#' }
#'
#' @seealso
#' \code{\link{call_llm}}
#' @export
NULL
list_models <- function(provider = c("github","openai","groq",
"anthropic","deepseek","dashscope",
"gemini","grok","all"),
...) {
provider <- match.arg(tolower(provider),
c("github","openai","groq",
"anthropic","deepseek","dashscope",
"gemini","grok","all"))
fetch <- switch(
provider,
"openai" = get_openai_models,
"groq" = get_groq_models,
"anthropic" = get_anthropic_models,
"deepseek" = get_deepseek_models,
"dashscope" = get_dashscope_models,
"gemini" = get_gemini_models, # new
"grok" = get_grok_models, # new
"github" = function(...) get_all_github_models(...),
"all" = NULL
)
if (!is.null(fetch)) {
mods <- tryCatch(fetch(...), error = function(e) character())
if (length(mods) == 0)
message(sprintf("No model catalog returned for '%s'.", provider))
return(mods)
}
provs <- c("openai","groq","anthropic","deepseek",
"dashscope","github","gemini","grok") # included in “all”
setNames(lapply(provs, function(p) {
tryCatch(list_models(p, ...), error = function(e) character())
}), provs)
}
###############################################################################
# 6. Core chat-completion wrapper #
###############################################################################
#' Core chat - completion wrapper for multiple providers
#'
#' @title Unified chat - completion interface
#' @name call_llm
#' @description
#' A unified wrapper for several "OpenAI - compatible" chat - completion APIs
#' (OpenAI, Groq, Anthropic, DeepSeek, Alibaba DashScope, GitHub Models, Grok, Gemini).
#' Accepts either a single `prompt` **or** a full `messages` list, adds the
#' correct authentication headers, retries on transient failures, and returns
#' the assistant's text response. You can toggle informational console
#' output with `verbose = TRUE/FALSE`. If the chosen `model` is no longer
#' available, the function stops early and suggests running
#' `list_models("<provider>")`.
#'
#' @section Messages:
#' * `prompt` - character scalar treated as a single *user* message.
#' * `messages` - list of lists; each element must contain `role` and `content`.
#' If both arguments are supplied, the `prompt` is appended
#' as an extra user message.
#'
#' @param prompt Character. Single user prompt (optional if `messages`).
#' @param messages List. Full chat history; see *Messages*.
#' @param provider Character. One of `"openai"`, `"groq"`, `"anthropic"`,
#' `"deepseek"`, `"dashscope"`,`"grok"`, `"gemini"` or `"github"`.
#' @param model Character. Model ID. If `NULL`, uses the provider default.
#' @param temperature Numeric. Sampling temperature (0 - 2). Default `0.7`.
#' @param max_tokens Integer. Max tokens to generate. Default `1000`.
#' @param api_key Character. Override API key; if `NULL`, uses the
#' environment variable for that provider.
#' @param n_tries Integer. Retry attempts on failure. Default `3`.
#' @param backoff Numeric. Seconds between retries. Default `2`.
#' @param verbose Logical. Whether to display informational messages
#' (`TRUE`) or suppress them (`FALSE`). Default `TRUE`.
#' @param endpoint_url Character. Custom endpoint; if `NULL`, a sensible
#' provider - specific default is used.
#' @param github_api_version Character. Header `X - GitHub - Api - Version`.
#' Default `"2022 - 11 - 28"`.
#' @param anthropic_api_version Character. Header `anthropic - version`.
#' Default `"2023 - 06 - 01"`.
#' @param ... Extra JSON - body fields (e.g. `top_p`, `stop`,
#' `presence_penalty`).
#' @param .post_func Internal. HTTP POST function (default `httr::POST`).
#'
#' @return Character scalar: assistant reply text.
#'
#' @examples
#' \dontrun{
#'
#' ## 1. Listing available models
#' # List all providers at once
#' all_mods <- list_models("all")
#' str(all_mods)
#'
#' # List OpenAI-only, Groq-only, Anthropic-only
#' openai_mods <- list_models("openai")
#' groq_mods <- list_models("groq")
#' anthropic_mods<- list_models("anthropic", anthropic_api_version = "2023-06-01")
#'
#' ## 2. Single-prompt interface
#'
#' # 2a. Basic usage
#' Sys.setenv(OPENAI_API_KEY = "sk-...")
#' res_basic <- call_llm(
#' prompt = "Hello, how are you?",
#' provider = "openai"
#' )
#' cat(res_basic)
#'
#' # 2b. Adjust sampling and penalties
#' res_sampling <- call_llm(
#' prompt = "Write a haiku about winter",
#' provider = "openai",
#' temperature = 1.2,
#' top_p = 0.5,
#' presence_penalty = 0.6,
#' frequency_penalty = 0.4
#' )
#' cat(res_sampling)
#'
#' # 2c. Control length and retries
#' res_len <- call_llm(
#' prompt = "List 5 uses for R",
#' provider = "openai",
#' max_tokens = 50,
#' n_tries = 5,
#' backoff = 0.5
#' )
#' cat(res_len)
#'
#' # 2d. Using stop sequences
#' res_stop <- call_llm(
#' prompt = "Count from 1 to 10:",
#' provider = "openai",
#' stop = c("6")
#' )
#' cat(res_stop)
#'
#' # 2e. Override API key for one call
#' res_override <- call_llm(
#' prompt = "Override test",
#' provider = "openai",
#' api_key = "sk-override",
#' max_tokens = 20
#' )
#' cat(res_override)
#'
#' # 2f. Factory interface for repeated prompts
#' GitHubLLM <- call_llm(
#' provider = "github",
#' max_tokens = 60,
#' verbose = FALSE
#' )
#' # direct invocation
#' story1 <- GitHubLLM("Tell me a short story")
#' cat(story1)
#'
#' ## 3. Multi-message conversation
#'
#' # 3a. Simple system + user
#' convo1 <- list(
#' list(role = "system", content = "You are a helpful assistant."),
#' list(role = "user", content = "Explain recursion.")
#' )
#' res1 <- call_llm(
#' messages = convo1,
#' provider = "openai",
#' max_tokens = 100
#' )
#' cat(res1)
#'
#' # 3b. Continue an existing chat by appending a prompt
#' prev <- list(
#' list(role = "system", content = "You are concise."),
#' list(role = "user", content = "Summarize the plot of Hamlet.")
#' )
#' res2 <- call_llm(
#' messages = prev,
#' prompt = "Now give me three bullet points."
#' )
#' cat(res2)
#'
#' # 3c. Use stop sequence in multi-message
#' convo2 <- list(
#' list(role = "system", content = "You list items."),
#' list(role = "user", content = "Name three colors.")
#' )
#' res3 <- call_llm(
#' messages = convo2,
#' stop = c(".")
#' )
#' cat(res3)
#'
#' # 3d. Multi-message via factory interface
#' ScopedLLM <- call_llm(provider = "openai", temperature = 0.3)
#' chat_ctx <- list(
#' list(role = "system", content = "You are a math tutor.")
#' )
#' ans1 <- ScopedLLM(messages = chat_ctx, prompt = "Solve 2+2.")
#' cat(ans1)
#' ans2 <- ScopedLLM("What about 10*10?")
#' cat(ans2)
#' }
#'
#' @export
NULL
call_llm <- function(
prompt = NULL,
messages = NULL,
provider = c("openai","groq","anthropic",
"deepseek","dashscope","github",
"gemini","grok"),
model = NULL,
temperature = 0.7,
max_tokens = 1000,
api_key = NULL,
n_tries = 3,
backoff = 2,
verbose = TRUE,
endpoint_url = NULL,
github_api_version = "2022-11-28",
anthropic_api_version = "2023-06-01",
...,
.post_func = httr::POST
) {
## ------------------------------------------------------------------------- ##
## Factory mode: if neither prompt nor messages supplied, return an LLM object
if (missing(prompt) && missing(messages)) {
return(
function(prompt = NULL,
messages = NULL,
...) {
# Re - invoke call_llm() with stored defaults + whatever the user
# passes now
args_main <- list(prompt = prompt, messages = messages)
opts_main <- list(
provider = provider,
model = model,
temperature = temperature,
max_tokens = max_tokens,
api_key = api_key,
n_tries = n_tries,
backoff = backoff,
verbose = verbose,
endpoint_url = endpoint_url,
github_api_version = github_api_version,
anthropic_api_version = anthropic_api_version,
.post_func = .post_func
)
extra_args <- list(...)
all_args <- c(args_main, opts_main, extra_args)
do.call(call_llm, all_args)
}
)
}
##
## ------------------------------------------------------------------------- ##
provider <- match.arg(tolower(provider),
c("openai","groq","anthropic",
"deepseek","dashscope","github",
"gemini","grok"))
if (is.null(model)) model <- get_default_model(provider)
## ---------------- assemble messages ----------------------------------- ##
if (!is.null(messages)) {
if (!is.null(prompt))
messages <- c(messages, list(list(role = "user", content = prompt)))
} else {
if (is.null(prompt))
stop("Provide either `prompt` or `messages`.")
messages <- list(list(role = "user", content = prompt))
}
## ---------------- common request pieces ------------------------------- ##
api_key <- get_api_key(provider, api_key)
req_body <- c(
list(
model = model,
messages = messages,
temperature = temperature,
max_tokens = max_tokens
),
list(...)
)
req_headers <- switch(
provider,
"openai" = add_headers(Authorization = paste("Bearer", api_key)),
"groq" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json"),
"anthropic" = add_headers(`x-api-key` = api_key,
`anthropic-version` = anthropic_api_version,
"Content-Type" = "application/json"),
"deepseek" = add_headers(Authorization = paste("Bearer", api_key)),
"dashscope" = add_headers(Authorization = paste("Bearer", api_key)),
"github" = add_headers(Accept = "application/vnd.github+json",
Authorization = paste("Bearer", api_key),
`X-GitHub-Api-Version` = github_api_version,
"Content-Type" = "application/json"),
"gemini" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json"),
"grok" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json")
)
if (is.null(endpoint_url)) {
endpoint_url <- switch(
provider,
"openai" = "https://api.openai.com/v1/chat/completions",
"groq" = "https://api.groq.com/openai/v1/chat/completions",
"anthropic" = "https://api.anthropic.com/v1/messages",
"deepseek" = "https://api.deepseek.com/v1/chat/completions",
"dashscope" = "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
"github" = {
org <- Sys.getenv("GH_MODELS_ORG")
if (nzchar(org))
sprintf("https://models.github.ai/orgs/%s/inference/chat/completions", org)
else
"https://models.github.ai/inference/chat/completions"
},
"gemini" = "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
"grok" = "https://api.x.ai/v1/chat/completions"
)
}
if (verbose) {
message(sprintf("Calling %s [%s] ... attempts=%d", provider, model, n_tries))
}
## ---------------- retry loop ------------------------------------------ ##
res <- NULL
for (i in seq_len(n_tries)) {
res <- tryCatch(
.post_func(
url = endpoint_url,
encode = "json",
body = req_body,
req_headers
),
error = function(e) {
# detect network time-outs explicitly
if (grepl("timeout", e$message, ignore.case = TRUE)) {
msg <- sprintf("Attempt %d/%d failed (%s: timeout). Retrying in %ds...",
i, n_tries, provider, backoff)
} else {
msg <- sprintf("Attempt %d/%d failed (%s). Retrying in %ds...",
i, n_tries, provider, backoff)
}
message(msg)
if (i == n_tries) {
stop(sprintf(
paste0(
'The request to provider "%s" timed out after %d attempt(s).\n\n',
'Tip: the model may be retired or misspelled.\n',
'Run list_models("%s") (after setting the proper *_API_KEY / ',
'MODELS_TOKEN) to see current models, e.g.\n',
' openai_models <- list_models("openai")\n',
'Then rerun call_llm(..., model = "<new-model>").\n\n',
'If the issue is network-related you can also:\n',
' . increase `n_tries` or `backoff`,\n',
' . provide a longer `timeout()` via `.post_func`, or\n',
' . check your network / VPN.\n\n',
'Internal message: %s'
),
provider, n_tries, provider, e$message
), call. = FALSE)
}
Sys.sleep(backoff)
NULL
}
)
if (is.null(res)) next
if (!http_error(res)) break
## ---- HTTP error branch --------------------------------------------- ##
err_txt <- content(res, "text", encoding = "UTF-8")
err_parsed <- tryCatch(fromJSON(err_txt), error = function(e) NULL)
not_found <- FALSE
if (!is.null(err_parsed$error$code))
not_found <- grepl("model_not_found|invalid_model|404",
err_parsed$error$code, ignore.case = TRUE)
if (!not_found && !is.null(err_parsed$message))
not_found <- grepl("model.*not.*found|no such model|de.?commiss|deprecated",
err_parsed$message, ignore.case = TRUE)
if (not_found) {
stop(sprintf(
paste0(
'The model "%s" is unavailable or de-commissioned for provider "%s".\n',
'Tip: run list_models("%s") after setting the proper *_API_KEY / ',
'MODELS_TOKEN to see current models, e.g.\n',
' openai_models <- list_models("openai")\n',
'Then rerun call_llm(..., model = "<new-model>").'
),
model, provider, provider
), call. = FALSE)
}
## --- generic HTTP error after final retry --------------------------- ##
if (i == n_tries) {
stop(sprintf(
paste0(
'Provider "%s" still returned an error after %d attempt(s).\n\n',
'Raw response from server:\n%s\n\n',
'Tip: the model may be retired, renamed, or misspelled.\n',
'. Run list_models("%s") (after setting the proper *_API_KEY / ',
'MODELS_TOKEN) to view currently available models, e.g.\n',
' openai_models <- list_models("openai")\n',
'. Or visit the provider\'s dashboard / documentation for the ',
'latest list.\n\n',
'Then rerun call_llm(..., model = "<new-model>").'
),
provider, n_tries, err_txt, provider
), call. = FALSE)
}
if (verbose) {
message(sprintf("HTTP %d on attempt %d/%d. Retrying in %ds...",
status_code(res), i, n_tries, backoff))
}
Sys.sleep(backoff)
}
txt <- parse_response(provider, content(res, "parsed"))
if (verbose) {
message(sprintf("Response (truncated):\n%s",
substr(txt, 1, min(200, nchar(txt)))))
}
txt
}
###############################################################################
# End of block #
###############################################################################
Try the chatLLM package in your browser
Any scripts or data that you put into this service are public.
chatLLM documentation built on Aug. 21, 2025, 6:03 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions call_llm list_models get_all_github_models get_dashscope_models get_deepseek_models get_anthropic_models get_groq_models get_gemini_models get_grok_models get_openai_models parse_response get_api_key get_default_model
Documented in call_llm list_models
###############################################################################
# chatLLM helpers + call_llm() #
# Requires: httr (>= 1.4.6), jsonlite #
###############################################################################
# ------------------------------------------------------------------------------
#' @importFrom httr POST add_headers http_error content
#' @importFrom jsonlite toJSON fromJSON
#' @importFrom httr GET timeout status_code
#' @importFrom stats setNames
library(httr)
library(jsonlite)
###############################################################################
# 1. Provider defaults #
###############################################################################
get_default_model <- function(provider) {
switch(
tolower(provider),
"openai" = "gpt-3.5-turbo",
"groq" = "meta-llama/llama-4-scout-17b-16e-instruct",
"anthropic" = "claude-3-7-sonnet-20250219",
"deepseek" = "deepseek-chat",
"dashscope" = "qwen-plus-latest",
"github" = "openai/gpt-4.1",
"gemini" = "gemini-2.0-flash",
"grok" = "grok-3-latest",
stop("No default model for provider: ", provider)
)
}
###############################################################################
# 2. API-key helper #
###############################################################################
get_api_key <- function(provider, api_key = NULL) {
env_var <- switch(
tolower(provider),
"openai" = "OPENAI_API_KEY",
"groq" = "GROQ_API_KEY",
"anthropic" = "ANTHROPIC_API_KEY",
"deepseek" = "DEEPSEEK_API_KEY",
"dashscope" = "DASHSCOPE_API_KEY",
"github" = "GH_MODELS_TOKEN",
"gemini" = "GEMINI_API_KEY",
"grok" = "XAI_API_KEY",
stop("Unknown provider: ", provider)
)
if (is.null(api_key)) api_key <- Sys.getenv(env_var)
if (!nzchar(api_key))
stop(sprintf("API key not found for %s. Set %s or pass `api_key`.",
provider, env_var))
api_key
}
###############################################################################
# 3. Parse chat-completion responses #
###############################################################################
parse_response <- function(provider, parsed) {
switch(
tolower(provider),
"openai" = parsed$choices[[1]]$message$content,
"groq" = parsed$choices[[1]]$message$content,
"anthropic" = parsed$content[[1]]$text,
"deepseek" = parsed$choices[[1]]$message$content,
"dashscope" = parsed$choices[[1]]$message$content,
"github" = parsed$choices[[1]]$message$content,
"gemini" = parsed$choices[[1]]$message$content,
"grok" = parsed$choices[[1]]$message$content,
stop("Parsing not implemented for provider: ", provider)
)
}
###############################################################################
# 4. Model-catalog helpers (one per provider) #
###############################################################################
get_openai_models <- function(token = Sys.getenv("OPENAI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.openai.com/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_grok_models <- function(token = Sys.getenv("XAI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.x.ai/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_gemini_models <- function(token = Sys.getenv("GEMINI_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://generativelanguage.googleapis.com/v1beta/openai/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_groq_models <- function(token = Sys.getenv("GROQ_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.groq.com/openai/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_anthropic_models <- function(token = Sys.getenv("ANTHROPIC_API_KEY"),
anthropic_api_version = "2023-06-01",
limit = 1000) {
if (!nzchar(token)) return(character())
r <- tryCatch(
GET("https://api.anthropic.com/v1/models",
add_headers(`x-api-key` = token,
`anthropic-version` = anthropic_api_version),
query = list(limit = limit),
timeout(60)),
error = function(e) NULL)
if (is.null(r) || http_error(r)) return(character())
p <- content(r, "parsed")
if (!is.null(p$data))
return(vapply(p$data, `[[`, character(1), "id"))
if (!is.null(p$models))
return(vapply(p$models, `[[`, character(1), "name"))
character()
}
get_deepseek_models <- function(token = Sys.getenv("DEEPSEEK_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://api.deepseek.com/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_dashscope_models <- function(token = Sys.getenv("DASHSCOPE_API_KEY")) {
if (!nzchar(token)) return(character())
r <- GET("https://dashscope-intl.aliyuncs.com/compatible-mode/v1/models",
add_headers(Authorization = paste("Bearer", token)),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed")$data, `[[`, character(1), "id")
}
get_all_github_models <- function(token = Sys.getenv("GH_MODELS_TOKEN"),
api_version = "2022-11-28") {
if (!nzchar(token)) return(character())
r <- GET("https://models.github.ai/catalog/models",
add_headers(Accept = "application/vnd.github+json",
Authorization = paste("Bearer", token),
`X-GitHub-Api-Version` = api_version),
timeout(60))
if (http_error(r)) return(character())
vapply(content(r, "parsed", simplifyVector = FALSE),
`[[`, character(1), "id")
}
###############################################################################
# 5. list_models() #
###############################################################################
#' List Available Models for Supported Providers
#'
#' @name list_models
#'
#' @description
#' Retrieve the catalog of available model IDs for one or all supported
#' chat - completion providers. Useful for discovering active models and
#' avoiding typos or deprecated defaults.
#'
#' Supported providers:
#' \itemize{
#' \item \code{"openai"} - OpenAI Chat Completions API
#' \item \code{"groq"} - Groq OpenAI - compatible endpoint
#' \item \code{"anthropic"} - Anthropic Claude API
#' \item \code{"deepseek"} - DeepSeek chat API
#' \item \code{"dashscope"} - Alibaba DashScope compatible API
#' \item \code{"github"} - GitHub Models OpenAI - compatible API
#' \item \code{"all"} - Fetch catalogs for all of the above
#' }
#'
#' @param provider Character. One of \code{"github"}, \code{"openai"},
#' \code{"groq"}, \code{"anthropic"}, \code{"deepseek"},
#' \code{"dashscope"} or \code{"all"}. Case - insensitive.
#' @param ... Additional arguments passed to the per - provider helper
#' (e.g. \code{limit} for Anthropic, or \code{api_version} for GitHub).
#' @param github_api_version Character. Header value for
#' \code{X - GitHub - Api - Version} (GitHub Models). Default \code{"2022 - 11 - 28"}.
#' @param anthropic_api_version Character. Header value for
#' \code{anthropic - version} (Anthropic). Default \code{"2023 - 06 - 01"}.
#'
#' @return
#' If \code{provider != "all"}, a character vector of model IDs for that
#' single provider. If \code{provider == "all"}, a named list of character
#' vectors, one per provider.
#'
#' @examples
#' \dontrun{
#' Sys.setenv(OPENAI_API_KEY = "sk-...")
#' openai_models <- list_models("openai")
#' head(openai_models)
#'
#' Sys.setenv(ANTHROPIC_API_KEY = "sk-...")
#' anthro_models <- list_models("anthropic", anthropic_api_version = "2023-06-01")
#'
#' Sys.setenv(GH_MODELS_TOKEN = "ghp-...")
#' github_models <- list_models("github", github_api_version = "2022-11-28")
#' }
#'
#' @seealso
#' \code{\link{call_llm}}
#' @export
NULL
list_models <- function(provider = c("github","openai","groq",
"anthropic","deepseek","dashscope",
"gemini","grok","all"),
...) {
provider <- match.arg(tolower(provider),
c("github","openai","groq",
"anthropic","deepseek","dashscope",
"gemini","grok","all"))
fetch <- switch(
provider,
"openai" = get_openai_models,
"groq" = get_groq_models,
"anthropic" = get_anthropic_models,
"deepseek" = get_deepseek_models,
"dashscope" = get_dashscope_models,
"gemini" = get_gemini_models, # new
"grok" = get_grok_models, # new
"github" = function(...) get_all_github_models(...),
"all" = NULL
)
if (!is.null(fetch)) {
mods <- tryCatch(fetch(...), error = function(e) character())
if (length(mods) == 0)
message(sprintf("No model catalog returned for '%s'.", provider))
return(mods)
}
provs <- c("openai","groq","anthropic","deepseek",
"dashscope","github","gemini","grok") # included in “all”
setNames(lapply(provs, function(p) {
tryCatch(list_models(p, ...), error = function(e) character())
}), provs)
}
###############################################################################
# 6. Core chat-completion wrapper #
###############################################################################
#' Core chat - completion wrapper for multiple providers
#'
#' @title Unified chat - completion interface
#' @name call_llm
#' @description
#' A unified wrapper for several "OpenAI - compatible" chat - completion APIs
#' (OpenAI, Groq, Anthropic, DeepSeek, Alibaba DashScope, GitHub Models, Grok, Gemini).
#' Accepts either a single `prompt` **or** a full `messages` list, adds the
#' correct authentication headers, retries on transient failures, and returns
#' the assistant's text response. You can toggle informational console
#' output with `verbose = TRUE/FALSE`. If the chosen `model` is no longer
#' available, the function stops early and suggests running
#' `list_models("<provider>")`.
#'
#' @section Messages:
#' * `prompt` - character scalar treated as a single *user* message.
#' * `messages` - list of lists; each element must contain `role` and `content`.
#' If both arguments are supplied, the `prompt` is appended
#' as an extra user message.
#'
#' @param prompt Character. Single user prompt (optional if `messages`).
#' @param messages List. Full chat history; see *Messages*.
#' @param provider Character. One of `"openai"`, `"groq"`, `"anthropic"`,
#' `"deepseek"`, `"dashscope"`,`"grok"`, `"gemini"` or `"github"`.
#' @param model Character. Model ID. If `NULL`, uses the provider default.
#' @param temperature Numeric. Sampling temperature (0 - 2). Default `0.7`.
#' @param max_tokens Integer. Max tokens to generate. Default `1000`.
#' @param api_key Character. Override API key; if `NULL`, uses the
#' environment variable for that provider.
#' @param n_tries Integer. Retry attempts on failure. Default `3`.
#' @param backoff Numeric. Seconds between retries. Default `2`.
#' @param verbose Logical. Whether to display informational messages
#' (`TRUE`) or suppress them (`FALSE`). Default `TRUE`.
#' @param endpoint_url Character. Custom endpoint; if `NULL`, a sensible
#' provider - specific default is used.
#' @param github_api_version Character. Header `X - GitHub - Api - Version`.
#' Default `"2022 - 11 - 28"`.
#' @param anthropic_api_version Character. Header `anthropic - version`.
#' Default `"2023 - 06 - 01"`.
#' @param ... Extra JSON - body fields (e.g. `top_p`, `stop`,
#' `presence_penalty`).
#' @param .post_func Internal. HTTP POST function (default `httr::POST`).
#'
#' @return Character scalar: assistant reply text.
#'
#' @examples
#' \dontrun{
#'
#' ## 1. Listing available models
#' # List all providers at once
#' all_mods <- list_models("all")
#' str(all_mods)
#'
#' # List OpenAI-only, Groq-only, Anthropic-only
#' openai_mods <- list_models("openai")
#' groq_mods <- list_models("groq")
#' anthropic_mods<- list_models("anthropic", anthropic_api_version = "2023-06-01")
#'
#' ## 2. Single-prompt interface
#'
#' # 2a. Basic usage
#' Sys.setenv(OPENAI_API_KEY = "sk-...")
#' res_basic <- call_llm(
#' prompt = "Hello, how are you?",
#' provider = "openai"
#' )
#' cat(res_basic)
#'
#' # 2b. Adjust sampling and penalties
#' res_sampling <- call_llm(
#' prompt = "Write a haiku about winter",
#' provider = "openai",
#' temperature = 1.2,
#' top_p = 0.5,
#' presence_penalty = 0.6,
#' frequency_penalty = 0.4
#' )
#' cat(res_sampling)
#'
#' # 2c. Control length and retries
#' res_len <- call_llm(
#' prompt = "List 5 uses for R",
#' provider = "openai",
#' max_tokens = 50,
#' n_tries = 5,
#' backoff = 0.5
#' )
#' cat(res_len)
#'
#' # 2d. Using stop sequences
#' res_stop <- call_llm(
#' prompt = "Count from 1 to 10:",
#' provider = "openai",
#' stop = c("6")
#' )
#' cat(res_stop)
#'
#' # 2e. Override API key for one call
#' res_override <- call_llm(
#' prompt = "Override test",
#' provider = "openai",
#' api_key = "sk-override",
#' max_tokens = 20
#' )
#' cat(res_override)
#'
#' # 2f. Factory interface for repeated prompts
#' GitHubLLM <- call_llm(
#' provider = "github",
#' max_tokens = 60,
#' verbose = FALSE
#' )
#' # direct invocation
#' story1 <- GitHubLLM("Tell me a short story")
#' cat(story1)
#'
#' ## 3. Multi-message conversation
#'
#' # 3a. Simple system + user
#' convo1 <- list(
#' list(role = "system", content = "You are a helpful assistant."),
#' list(role = "user", content = "Explain recursion.")
#' )
#' res1 <- call_llm(
#' messages = convo1,
#' provider = "openai",
#' max_tokens = 100
#' )
#' cat(res1)
#'
#' # 3b. Continue an existing chat by appending a prompt
#' prev <- list(
#' list(role = "system", content = "You are concise."),
#' list(role = "user", content = "Summarize the plot of Hamlet.")
#' )
#' res2 <- call_llm(
#' messages = prev,
#' prompt = "Now give me three bullet points."
#' )
#' cat(res2)
#'
#' # 3c. Use stop sequence in multi-message
#' convo2 <- list(
#' list(role = "system", content = "You list items."),
#' list(role = "user", content = "Name three colors.")
#' )
#' res3 <- call_llm(
#' messages = convo2,
#' stop = c(".")
#' )
#' cat(res3)
#'
#' # 3d. Multi-message via factory interface
#' ScopedLLM <- call_llm(provider = "openai", temperature = 0.3)
#' chat_ctx <- list(
#' list(role = "system", content = "You are a math tutor.")
#' )
#' ans1 <- ScopedLLM(messages = chat_ctx, prompt = "Solve 2+2.")
#' cat(ans1)
#' ans2 <- ScopedLLM("What about 10*10?")
#' cat(ans2)
#' }
#'
#' @export
NULL
call_llm <- function(
prompt = NULL,
messages = NULL,
provider = c("openai","groq","anthropic",
"deepseek","dashscope","github",
"gemini","grok"),
model = NULL,
temperature = 0.7,
max_tokens = 1000,
api_key = NULL,
n_tries = 3,
backoff = 2,
verbose = TRUE,
endpoint_url = NULL,
github_api_version = "2022-11-28",
anthropic_api_version = "2023-06-01",
...,
.post_func = httr::POST
) {
## ------------------------------------------------------------------------- ##
## Factory mode: if neither prompt nor messages supplied, return an LLM object
if (missing(prompt) && missing(messages)) {
return(
function(prompt = NULL,
messages = NULL,
...) {
# Re - invoke call_llm() with stored defaults + whatever the user
# passes now
args_main <- list(prompt = prompt, messages = messages)
opts_main <- list(
provider = provider,
model = model,
temperature = temperature,
max_tokens = max_tokens,
api_key = api_key,
n_tries = n_tries,
backoff = backoff,
verbose = verbose,
endpoint_url = endpoint_url,
github_api_version = github_api_version,
anthropic_api_version = anthropic_api_version,
.post_func = .post_func
)
extra_args <- list(...)
all_args <- c(args_main, opts_main, extra_args)
do.call(call_llm, all_args)
}
)
}
##
## ------------------------------------------------------------------------- ##
provider <- match.arg(tolower(provider),
c("openai","groq","anthropic",
"deepseek","dashscope","github",
"gemini","grok"))
if (is.null(model)) model <- get_default_model(provider)
## ---------------- assemble messages ----------------------------------- ##
if (!is.null(messages)) {
if (!is.null(prompt))
messages <- c(messages, list(list(role = "user", content = prompt)))
} else {
if (is.null(prompt))
stop("Provide either `prompt` or `messages`.")
messages <- list(list(role = "user", content = prompt))
}
## ---------------- common request pieces ------------------------------- ##
api_key <- get_api_key(provider, api_key)
req_body <- c(
list(
model = model,
messages = messages,
temperature = temperature,
max_tokens = max_tokens
),
list(...)
)
req_headers <- switch(
provider,
"openai" = add_headers(Authorization = paste("Bearer", api_key)),
"groq" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json"),
"anthropic" = add_headers(`x-api-key` = api_key,
`anthropic-version` = anthropic_api_version,
"Content-Type" = "application/json"),
"deepseek" = add_headers(Authorization = paste("Bearer", api_key)),
"dashscope" = add_headers(Authorization = paste("Bearer", api_key)),
"github" = add_headers(Accept = "application/vnd.github+json",
Authorization = paste("Bearer", api_key),
`X-GitHub-Api-Version` = github_api_version,
"Content-Type" = "application/json"),
"gemini" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json"),
"grok" = add_headers(Authorization = paste("Bearer", api_key),
"Content-Type" = "application/json")
)
if (is.null(endpoint_url)) {
endpoint_url <- switch(
provider,
"openai" = "https://api.openai.com/v1/chat/completions",
"groq" = "https://api.groq.com/openai/v1/chat/completions",
"anthropic" = "https://api.anthropic.com/v1/messages",
"deepseek" = "https://api.deepseek.com/v1/chat/completions",
"dashscope" = "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions",
"github" = {
org <- Sys.getenv("GH_MODELS_ORG")
if (nzchar(org))
sprintf("https://models.github.ai/orgs/%s/inference/chat/completions", org)
else
"https://models.github.ai/inference/chat/completions"
},
"gemini" = "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
"grok" = "https://api.x.ai/v1/chat/completions"
)
}
if (verbose) {
message(sprintf("Calling %s [%s] ... attempts=%d", provider, model, n_tries))
}
## ---------------- retry loop ------------------------------------------ ##
res <- NULL
for (i in seq_len(n_tries)) {
res <- tryCatch(
.post_func(
url = endpoint_url,
encode = "json",
body = req_body,
req_headers
),
error = function(e) {
# detect network time-outs explicitly
if (grepl("timeout", e$message, ignore.case = TRUE)) {
msg <- sprintf("Attempt %d/%d failed (%s: timeout). Retrying in %ds...",
i, n_tries, provider, backoff)
} else {
msg <- sprintf("Attempt %d/%d failed (%s). Retrying in %ds...",
i, n_tries, provider, backoff)
}
message(msg)
if (i == n_tries) {
stop(sprintf(
paste0(
'The request to provider "%s" timed out after %d attempt(s).\n\n',
'Tip: the model may be retired or misspelled.\n',
'Run list_models("%s") (after setting the proper *_API_KEY / ',
'MODELS_TOKEN) to see current models, e.g.\n',
' openai_models <- list_models("openai")\n',
'Then rerun call_llm(..., model = "<new-model>").\n\n',
'If the issue is network-related you can also:\n',
' . increase `n_tries` or `backoff`,\n',
' . provide a longer `timeout()` via `.post_func`, or\n',
' . check your network / VPN.\n\n',
'Internal message: %s'
),
provider, n_tries, provider, e$message
), call. = FALSE)
}
Sys.sleep(backoff)
NULL
}
)
if (is.null(res)) next
if (!http_error(res)) break
## ---- HTTP error branch --------------------------------------------- ##
err_txt <- content(res, "text", encoding = "UTF-8")
err_parsed <- tryCatch(fromJSON(err_txt), error = function(e) NULL)
not_found <- FALSE
if (!is.null(err_parsed$error$code))
not_found <- grepl("model_not_found|invalid_model|404",
err_parsed$error$code, ignore.case = TRUE)
if (!not_found && !is.null(err_parsed$message))
not_found <- grepl("model.*not.*found|no such model|de.?commiss|deprecated",
err_parsed$message, ignore.case = TRUE)
if (not_found) {
stop(sprintf(
paste0(
'The model "%s" is unavailable or de-commissioned for provider "%s".\n',
'Tip: run list_models("%s") after setting the proper *_API_KEY / ',
'MODELS_TOKEN to see current models, e.g.\n',
' openai_models <- list_models("openai")\n',
'Then rerun call_llm(..., model = "<new-model>").'
),
model, provider, provider
), call. = FALSE)
}
## --- generic HTTP error after final retry --------------------------- ##
if (i == n_tries) {
stop(sprintf(
paste0(
'Provider "%s" still returned an error after %d attempt(s).\n\n',
'Raw response from server:\n%s\n\n',
'Tip: the model may be retired, renamed, or misspelled.\n',
'. Run list_models("%s") (after setting the proper *_API_KEY / ',
'MODELS_TOKEN) to view currently available models, e.g.\n',
' openai_models <- list_models("openai")\n',
'. Or visit the provider\'s dashboard / documentation for the ',
'latest list.\n\n',
'Then rerun call_llm(..., model = "<new-model>").'
),
provider, n_tries, err_txt, provider
), call. = FALSE)
}
if (verbose) {
message(sprintf("HTTP %d on attempt %d/%d. Retrying in %ds...",
status_code(res), i, n_tries, backoff))
}
Sys.sleep(backoff)
}
txt <- parse_response(provider, content(res, "parsed"))
if (verbose) {
message(sprintf("Response (truncated):\n%s",
substr(txt, 1, min(200, nchar(txt)))))
}
txt
}
###############################################################################
# End of block #
###############################################################################
Try the chatLLM package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.