Package 'inferencer'

Title: Simple Unified Wrappers for Foundation Model Inference APIs
Description: Provides lightweight R wrappers for querying and listing models from several hosted and local foundation model inference platforms, currently including 'OpenAI' <https://developers.openai.com/api/docs/>, 'Google Gemini' <https://ai.google.dev/gemini-api/docs>, 'Groq' <https://console.groq.com/docs/>, 'OpenRouter' <https://openrouter.ai/docs/>, 'Cerebras' <https://inference-docs.cerebras.ai/>, 'Ollama Cloud' <https://docs.ollama.com/cloud>, 'Alibaba Cloud Model Studio' <https://help.aliyun.com/en/model-studio/>, 'Zhipu AI' <https://docs.bigmodel.cn/>, 'DeepSeek' <https://api-docs.deepseek.com/>, 'Moonshot AI' <https://platform.kimi.ai/docs/>, and 'MiniMax' <https://platform.minimax.io/docs/>. The package is designed for simple inference workflows and quick experimentation, with minimal abstraction and a consistent interface across providers. It includes helper functions for model discovery, text generation, embeddings, image generation, and multimodal inputs, while leaving room for future support of provider-specific parameters and advanced options.
Authors: Oliver Zhou [aut, cre]
Maintainer: Oliver Zhou <[email protected]>
License: MIT + file LICENSE
Version: 0.2.2
Built: 2026-10-01 07:47:11 UTC
Source: https://github.com/oliverlds/inferencer

Help Index

Request Gemini Embeddings

Description

Calls Gemini embedding endpoints for one or more text inputs.

Usage

embed_gemini(
  input,
  api_key = Sys.getenv("GEMINI_API_KEY"),
  model = "gemini-embedding-001",
  url0 = Sys.getenv("GEMINI_API_URL", unset =
    "https://generativelanguage.googleapis.com/v1beta/models"),
  task_type = NULL,
  title = NULL,
  output_dimensionality = NULL,
  json_list = FALSE
)

Arguments

input

A non-empty character vector.
api_key

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
model

Embedding model identifier.
url0

Base Gemini models URL.
task_type

Optional Gemini embedding taskType.
title

Optional embedding title, used with document-style inputs.
output_dimensionality

Optional embedding dimensionality override.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A numeric matrix by default, or the parsed JSON response when json_list = TRUE.

Request OpenRouter Embeddings

Description

Calls the OpenRouter embeddings endpoint for one or more text inputs.

Usage

embed_openrouter(
  input,
  model = "openai/text-embedding-3-small",
  dimensions = NULL,
  encoding_format = c("float", "base64"),
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/embeddings",
  json_list = FALSE
)

Arguments

input

A non-empty character vector.
model

Model identifier.
dimensions

Optional embedding dimensionality override.
encoding_format

Embedding encoding. Defaults to "float".
api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter embeddings endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A numeric matrix by default when encoding_format = "float", or the parsed JSON response when json_list = TRUE.

Extract Benchmark Scores from OpenRouter Model Metadata

Description

Scans OpenRouter model metadata for benchmark payloads, including possible Artificial Analysis benchmark fields when they are present in the models response.

Usage

extract_openrouter_benchmarks(
  models,
  benchmark_fields = c("benchmarks", "benchmark_scores", "artificial_analysis",
    "artificial_analysis_benchmarks", "benchmark_data")
)

Arguments

models

Parsed JSON from list_openrouter_models(json_list = TRUE), a list of OpenRouter model objects, or a data.frame returned by list_openrouter_models().
benchmark_fields

Candidate field names to inspect on each model.

Value

A data.table with one row per extracted benchmark metric. Returns an empty data.table when no benchmark fields are found.

Generate Images with Gemini

Description

Calls Gemini image-generation capable models and returns the generated image payloads as base64 strings.

Usage

generate_image_gemini(
  prompt,
  api_key = Sys.getenv("GEMINI_API_KEY"),
  model = "gemini-2.5-flash-image",
  url0 = Sys.getenv("GEMINI_API_URL", unset =
    "https://generativelanguage.googleapis.com/v1beta/models"),
  temperature = 0.7,
  top_p = 1,
  top_k = 40,
  max_tokens = NULL,
  response_modalities = c("TEXT", "IMAGE"),
  json_list = FALSE
)

Arguments

prompt

A non-empty character string.
api_key

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
model

Image-capable Gemini model identifier.
url0

Base Gemini models URL.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
top_k

Top-k sampling parameter.
max_tokens

Optional maximum number of output tokens.
response_modalities

Modalities to request. Defaults to c("TEXT", "IMAGE").
json_list

If TRUE, return the parsed JSON response as a list.

Value

A character vector of base64-encoded image payloads by default, or the parsed JSON response when json_list = TRUE.

Generate Images with OpenRouter

Description

Calls the OpenRouter chat completions API for image-capable models and returns the first available image payload or URL.

Usage

generate_image_openrouter(
  prompt,
  model = "google/gemini-2.5-flash-image-preview",
  temperature = 0,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = Sys.getenv("OPENROUTER_API_URL", unset =
    "https://openrouter.ai/api/v1/chat/completions"),
  json_list = FALSE
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A character string containing an image URL or base64 payload by default, or the parsed JSON response when json_list = TRUE.

List Cerebras Models

Description

Retrieves available models from the public Cerebras models endpoint.

Usage

list_cerebras_models(
  url = "https://api.cerebras.ai/public/v1/models",
  json_list = FALSE
)

Arguments

url

Cerebras public models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Verified No-Cost Mainland-China Models

Description

Returns models that Inferencer can verify as no-cost APIs rather than models with account-specific trials or expiring promotional quotas.

Usage

list_china_free_models()

Value

A data.table with provider, model, free_basis, and source_url.

List DeepSeek Models

Description

Retrieves models available from the DeepSeek API.

Usage

list_deepseek_models(
  api_key = Sys.getenv("DEEPSEEK_API_KEY"),
  url = "https://api.deepseek.com/models",
  json_list = FALSE,
  timeout = 120
)

Arguments

api_key

DeepSeek API key. Defaults to Sys.getenv("DEEPSEEK_API_KEY").
url

DeepSeek OpenAI-compatible models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Gemini Models

Description

Retrieves available models from the Gemini models endpoint.

Usage

list_gemini_models(
  api_key = Sys.getenv("GEMINI_API_KEY"),
  url = "https://generativelanguage.googleapis.com/v1beta/models",
  json_list = FALSE
)

Arguments

api_key

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
url

Gemini models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Groq Models

Description

Retrieves available models from the Groq models endpoint.

Usage

list_groq_models(
  api_key = Sys.getenv("GROQ_API_KEY"),
  url = "https://api.groq.com/openai/v1/models",
  json_list = FALSE
)

Arguments

api_key

Groq API key. Defaults to Sys.getenv("GROQ_API_KEY").
url

Groq models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List MiniMax Models

Description

Retrieves models available from the MiniMax API.

Usage

list_minimax_models(
  api_key = Sys.getenv("MINIMAX_API_KEY"),
  url = Sys.getenv("MINIMAX_MODELS_URL", unset = "https://api.minimax.io/v1/models"),
  json_list = FALSE,
  timeout = 120
)

Arguments

api_key

MiniMax API key. Defaults to Sys.getenv("MINIMAX_API_KEY").
url

MiniMax OpenAI-compatible models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Moonshot Models

Description

Retrieves models available from the Moonshot AI Kimi API.

Usage

list_moonshot_models(
  api_key = Sys.getenv("MOONSHOT_API_KEY"),
  url = "https://api.moonshot.ai/v1/models",
  json_list = FALSE,
  timeout = 120
)

Arguments

api_key

Moonshot API key. Defaults to Sys.getenv("MOONSHOT_API_KEY").
url

Moonshot OpenAI-compatible models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Ollama Cloud Models

Description

Retrieves available models from the Ollama Cloud tags endpoint.

Usage

list_ollama_models(
  api_key = Sys.getenv("OLLAMA_API_KEY"),
  url = "https://ollama.com/api/tags",
  json_list = FALSE
)

Arguments

api_key

Ollama API key. Defaults to Sys.getenv("OLLAMA_API_KEY").
url

Ollama Cloud tags endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Models from an OpenAI-Compatible Server

Description

Retrieves models from an OpenAI-compatible ⁠/v1/models⁠ endpoint. This is suitable for local or self-hosted servers such as llama.cpp, vLLM, LM Studio, LocalAI, and compatible deployments. Authentication is optional.

Usage

list_openai_compatible_models(
  url = "http://127.0.0.1:8080/v1/models",
  api_key = NULL,
  json_list = FALSE,
  timeout = 120
)

Arguments

url

OpenAI-compatible models endpoint.
api_key

Optional bearer token. Use NULL for an unauthenticated local server.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List OpenAI Models

Description

Retrieves models available to the current OpenAI API key.

Usage

list_openai_models(
  api_key = Sys.getenv("OPENAI_API_KEY"),
  url = "https://api.openai.com/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenAI API key. Defaults to Sys.getenv("OPENAI_API_KEY").
url

OpenAI models endpoint.
json_list

If TRUE, return the complete parsed JSON response as a list.

Value

A data.table by default, or the complete parsed JSON response when json_list = TRUE.

List OpenRouter Audio Models

Description

Filters the general OpenRouter models catalog to models with audio input or output modalities.

Usage

list_openrouter_audio_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter models endpoint.
json_list

If TRUE, return the filtered parsed JSON response as a list.

Value

A filtered data.table by default, or a filtered parsed JSON list when json_list = TRUE.

List OpenRouter Benchmarks

Description

Retrieves benchmark results directly from OpenRouter's unified benchmarks endpoint. Unlike extract_openrouter_benchmarks(), this function does not inspect model metadata and is intended for benchmark refresh workflows.

Usage

list_openrouter_benchmarks(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  source = NULL,
  task_type = NULL,
  url = "https://openrouter.ai/api/v1/benchmarks"
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
source

Optional benchmark source: "artificial-analysis", "design-arena", or "openrouter".
task_type

Optional workload filter: "intelligence", "coding", or "agentic".
url

OpenRouter benchmarks endpoint.

Value

A long data.table with one row per model and benchmark metric. The returned scores retain the source-native scale. Provider-supplied source timestamps and URLs, plus the queried endpoint, are retained for provenance. price_basis identifies provider-reported per-token pricing and is distinct from avg_cost_per_task.

List OpenRouter Embedding Models

Description

Filters the general OpenRouter models catalog to models with embedding output modalities.

Usage

list_openrouter_embedding_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter models endpoint.
json_list

If TRUE, return the filtered parsed JSON response as a list.

Value

A filtered data.table by default, or a filtered parsed JSON list when json_list = TRUE.

List OpenRouter Image Models

Description

Filters the general OpenRouter models catalog to models with image output modalities.

Usage

list_openrouter_image_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter models endpoint.
json_list

If TRUE, return the filtered parsed JSON response as a list.

Value

A filtered data.table by default, or a filtered parsed JSON list when json_list = TRUE.

List OpenRouter Models

Description

Retrieves available models from the OpenRouter models endpoint.

Usage

list_openrouter_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List OpenRouter Multimodal Models

Description

Filters the general OpenRouter models catalog to models that support multiple input modalities.

Usage

list_openrouter_multimodal_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter models endpoint.
json_list

If TRUE, return the filtered parsed JSON response as a list.

Value

A filtered data.table by default, or a filtered parsed JSON list when json_list = TRUE.

List OpenRouter Video Generation Models

Description

Retrieves available video generation models and their normalized feature metadata from the OpenRouter video models endpoint.

Usage

list_openrouter_video_models(
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = "https://openrouter.ai/api/v1/videos/models",
  json_list = FALSE
)

Arguments

api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter video models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Qwen Models

Description

Retrieves models available from Alibaba Cloud Model Studio.

Usage

list_qwen_models(
  api_key = Sys.getenv("DASHSCOPE_API_KEY"),
  url = Sys.getenv("QWEN_MODELS_URL", unset =
    "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/models"),
  json_list = FALSE,
  timeout = 120
)

Arguments

api_key

Qwen API key. Defaults to Sys.getenv("DASHSCOPE_API_KEY").
url

Qwen OpenAI-compatible models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

List Zhipu Models

Description

Retrieves models available from the Zhipu AI Open Platform.

Usage

list_zhipu_models(
  api_key = Sys.getenv("ZHIPUAI_API_KEY"),
  url = "https://open.bigmodel.cn/api/paas/v4/models",
  json_list = FALSE,
  timeout = 120
)

Arguments

api_key

Zhipu AI API key. Defaults to Sys.getenv("ZHIPUAI_API_KEY").
url

Zhipu OpenAI-compatible models endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A data.table by default, or a parsed JSON list when json_list = TRUE.

Query a Cerebras Chat Model

Description

Sends a single user prompt to the Cerebras chat completions API.

Usage

query_cerebras(
  prompt,
  model = "gpt-oss-120b",
  api_key = Sys.getenv("CEREBRAS_API_KEY"),
  url = "https://api.cerebras.ai/v1/chat/completions",
  json_list = FALSE
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
api_key

Cerebras API key. Defaults to Sys.getenv("CEREBRAS_API_KEY").
url

Cerebras chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A character string by default, or a parsed JSON list when json_list = TRUE.

See Also

list_cerebras_models()

Query Mainland-China Providers with Ordered Fallback

Description

Tries mainland-China-hosted provider APIs in order: Qwen, Zhipu, DeepSeek, Moonshot, then MiniMax.

Usage

query_china_fallback(
  prompt,
  json_list = FALSE,
  timeout = 120,
  api_key_qwen = Sys.getenv("DASHSCOPE_API_KEY"),
  api_key_zhipu = Sys.getenv("ZHIPUAI_API_KEY"),
  api_key_deepseek = Sys.getenv("DEEPSEEK_API_KEY"),
  api_key_moonshot = Sys.getenv("MOONSHOT_API_KEY"),
  api_key_minimax = Sys.getenv("MINIMAX_API_KEY")
)

Arguments

prompt

A non-empty character string.
json_list

If TRUE, return a list containing the successful provider name and parsed response object.
timeout

Maximum number of seconds to wait for each provider request.
api_key_qwen

Qwen API key. Defaults to Sys.getenv("DASHSCOPE_API_KEY").
api_key_zhipu

Zhipu AI API key. Defaults to Sys.getenv("ZHIPUAI_API_KEY").
api_key_deepseek

DeepSeek API key. Defaults to Sys.getenv("DEEPSEEK_API_KEY").
api_key_moonshot

Moonshot API key. Defaults to Sys.getenv("MOONSHOT_API_KEY").
api_key_minimax

MiniMax API key. Defaults to Sys.getenv("MINIMAX_API_KEY").

Value

A character string by default. When json_list = TRUE, returns a list with elements provider and response.

Query a Verified No-Cost Mainland-China Model

Description

Calls only the currently verified no-cost mainland-China model. This avoids treating account-specific free trials as universally free and therefore does not silently route a request to a paid model.

Usage

query_china_free_fallback(
  prompt,
  json_list = FALSE,
  timeout = 120,
  api_key_zhipu = Sys.getenv("ZHIPUAI_API_KEY")
)

Arguments

prompt

A non-empty character string.
json_list

If TRUE, return a list containing the successful provider name and parsed response object.
timeout

Maximum number of seconds to wait for the provider request.
api_key_zhipu

Zhipu AI API key. Defaults to Sys.getenv("ZHIPUAI_API_KEY").

Value

A character string by default. When json_list = TRUE, returns a list with elements provider and response.

See Also

list_china_free_models()

Query a DeepSeek Model

Description

Sends a prompt to the DeepSeek OpenAI-compatible chat endpoint.

Usage

query_deepseek(
  prompt,
  model = "deepseek-v4-flash",
  temperature = 0.7,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("DEEPSEEK_API_KEY"),
  url = "https://api.deepseek.com/chat/completions",
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

DeepSeek API key. Defaults to Sys.getenv("DEEPSEEK_API_KEY").
url

DeepSeek OpenAI-compatible chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

See Also

list_deepseek_models()

Query Providers with Ordered Fallback

Description

Tries the package's text-query wrappers in order until one succeeds: query_gemini(), then query_openrouter(), then query_groq().

Usage

query_fallback(
  prompt,
  json_list = FALSE,
  api_key_gemini = Sys.getenv("GEMINI_API_KEY"),
  api_key_openrouter = Sys.getenv("OPENROUTER_API_KEY"),
  api_key_groq = Sys.getenv("GROQ_API_KEY"),
  timeout = 120
)

Arguments

prompt

A non-empty character string.
json_list

If TRUE, return a list containing the successful provider name and parsed response object.
api_key_gemini

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
api_key_openrouter

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
api_key_groq

Groq API key. Defaults to Sys.getenv("GROQ_API_KEY").
timeout

Maximum number of seconds to wait for each provider request.

Value

A character string by default. When json_list = TRUE, returns a list with elements provider and response.

Query a Gemini Model

Description

Sends a single user prompt to the Gemini generateContent API.

Usage

query_gemini(
  prompt,
  api_key = Sys.getenv("GEMINI_API_KEY"),
  model = "gemini-2.5-flash",
  url0 = Sys.getenv("GEMINI_API_URL", unset =
    "https://generativelanguage.googleapis.com/v1beta/models"),
  temperature = 0.7,
  top_p = 1,
  top_k = 40,
  max_tokens = NULL,
  response_modalities = NULL,
  speech_config = NULL,
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
api_key

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
model

Model identifier.
url0

Base Gemini models URL.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
top_k

Top-k sampling parameter.
max_tokens

Optional maximum number of output tokens.
response_modalities

Optional response modalities, for example c("TEXT") or c("AUDIO").
speech_config

Optional Gemini speechConfig object supplied as an R list.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string by default. For audio responses, returns the base64-encoded audio data from inlineData$data. Returns the parsed JSON list when json_list = TRUE.

Query a Gemini Multimodal Model

Description

Sends Gemini generateContent requests using either a simple text prompt or an explicit list of content parts. This allows text, image, audio, PDF, and other supported non-text inputs to be passed through the same wrapper.

Usage

query_gemini_content(
  prompt = NULL,
  parts = NULL,
  api_key = Sys.getenv("GEMINI_API_KEY"),
  model = "gemini-2.5-flash",
  url0 = Sys.getenv("GEMINI_API_URL", unset =
    "https://generativelanguage.googleapis.com/v1beta/models"),
  temperature = 0.7,
  top_p = 1,
  top_k = 40,
  max_tokens = NULL,
  response_modalities = NULL,
  speech_config = NULL,
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string. Ignored when parts is supplied.
parts

Optional Gemini parts payload supplied as a list. Use this for multimodal inputs such as inlineData or fileData.
api_key

Gemini API key. Defaults to Sys.getenv("GEMINI_API_KEY").
model

Model identifier.
url0

Base Gemini models URL.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
top_k

Top-k sampling parameter.
max_tokens

Optional maximum number of output tokens.
response_modalities

Optional response modalities, for example c("TEXT"), c("AUDIO"), or c("TEXT", "IMAGE").
speech_config

Optional Gemini speechConfig object supplied as an R list.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string when Gemini returns text, a base64 string when the first returned part is inline binary data, or the parsed JSON response when json_list = TRUE.

Query a Groq Chat Model

Description

Sends a single user prompt to the Groq chat completions API.

Usage

query_groq(
  prompt,
  api_key = Sys.getenv("GROQ_API_KEY"),
  url = Sys.getenv("GROQ_API_URL", unset =
    "https://api.groq.com/openai/v1/chat/completions"),
  model = c("groq/compound", "allam-2-7b", "groq/compound-mini", "qwen/qwen3-32b",
    "openai/gpt-oss-20b", "canopylabs/orpheus-v1-english", "openai/gpt-oss-120b",
    "whisper-large-v3", "llama-3.3-70b-versatile", "moonshotai/kimi-k2-instruct-0905",
    "whisper-large-v3-turbo", "meta-llama/llama-prompt-guard-2-86m",
    "moonshotai/kimi-k2-instruct", "meta-llama/llama-prompt-guard-2-22m",
    "meta-llama/llama-4-scout-17b-16e-instruct", "openai/gpt-oss-safeguard-20b",
    "llama-3.1-8b-instant", "canopylabs/orpheus-arabic-saudi"),
  temperature = 1,
  top_p = 1,
  max_tokens = 1024,
  stream = FALSE,
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
api_key

Groq API key. Defaults to Sys.getenv("GROQ_API_KEY").
url

Groq chat completions endpoint.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
stream

Whether to request an SSE response. The complete response is buffered and assembled before the function returns; tokens are not printed incrementally to the terminal.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string by default. With json_list = TRUE, returns a parsed JSON response for non-streaming requests or a list of parsed SSE chunks for streaming requests.

Query a MiniMax Model

Description

Sends a prompt to the MiniMax OpenAI-compatible chat endpoint.

Usage

query_minimax(
  prompt,
  model = "MiniMax-M2.7",
  temperature = 0.7,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("MINIMAX_API_KEY"),
  url = Sys.getenv("MINIMAX_API_URL", unset =
    "https://api.minimax.io/v1/chat/completions"),
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

MiniMax API key. Defaults to Sys.getenv("MINIMAX_API_KEY").
url

MiniMax OpenAI-compatible chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

See Also

list_minimax_models()

Query a Moonshot Model

Description

Sends a prompt to the Moonshot AI Kimi OpenAI-compatible chat endpoint.

Usage

query_moonshot(
  prompt,
  model = "kimi-k2.6",
  temperature = 0.7,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("MOONSHOT_API_KEY"),
  url = "https://api.moonshot.ai/v1/chat/completions",
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

Moonshot API key. Defaults to Sys.getenv("MOONSHOT_API_KEY").
url

Moonshot OpenAI-compatible chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

See Also

list_moonshot_models()

Query an Ollama Cloud Chat Model

Description

Sends a single user prompt to the Ollama Cloud chat API.

Usage

query_ollama(
  prompt,
  model = "gpt-oss:120b",
  api_key = Sys.getenv("OLLAMA_API_KEY"),
  url = "https://ollama.com/api/chat",
  json_list = FALSE
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
api_key

Ollama API key. Defaults to Sys.getenv("OLLAMA_API_KEY").
url

Ollama Cloud chat endpoint.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A character string by default, or a parsed JSON list when json_list = TRUE.

Query an OpenAI Model with the Responses API

Description

Sends a request directly to OpenAI's Responses API. The input may be a simple text prompt or a structured Responses API input list for multimodal and multi-message workflows.

Usage

query_openai(
  input,
  model = "gpt-5.6",
  instructions = NULL,
  max_output_tokens = NULL,
  temperature = NULL,
  top_p = NULL,
  tools = NULL,
  api_key = Sys.getenv("OPENAI_API_KEY"),
  url = Sys.getenv("OPENAI_API_URL", unset = "https://api.openai.com/v1/responses"),
  json_list = FALSE
)

Arguments

input

A non-empty character string or structured Responses API input list.
model

Model identifier.
instructions

Optional system or developer instructions.
max_output_tokens

Optional maximum number of generated tokens, including visible output and reasoning tokens.
temperature

Optional sampling temperature.
top_p

Optional nucleus sampling parameter.
tools

Optional list of Responses API tool definitions.
api_key

OpenAI API key. Defaults to Sys.getenv("OPENAI_API_KEY").
url

OpenAI Responses endpoint.
json_list

If TRUE, return the complete parsed JSON response as a list.

Value

A character string containing the response text by default, or the complete parsed JSON response when json_list = TRUE.

Query an OpenAI-Compatible Chat Server

Description

Sends a prompt to an OpenAI-compatible ⁠/v1/chat/completions⁠ endpoint. This is suitable for local or self-hosted servers such as llama.cpp, vLLM, LM Studio, LocalAI, and compatible deployments. Authentication is optional. response_format is forwarded unchanged, allowing servers that support JSON-schema constrained output to use their native request shape.

Usage

query_openai_compatible(
  prompt,
  model,
  system = NULL,
  response_format = NULL,
  temperature = NULL,
  max_tokens = NULL,
  url = "http://127.0.0.1:8080/v1/chat/completions",
  api_key = NULL,
  timeout = 120,
  json_list = FALSE
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
system

Optional system prompt.
response_format

Optional OpenAI-compatible response-format object.
temperature

Optional sampling temperature.
max_tokens

Optional maximum number of output tokens.
url

OpenAI-compatible chat completions endpoint.
api_key

Optional bearer token. Use NULL for an unauthenticated local server.
timeout

Maximum number of seconds to wait for the HTTP request.
json_list

If TRUE, return the parsed JSON response as a list.

Value

A character string by default, or a parsed JSON list when json_list = TRUE.

Query an OpenRouter Chat Model

Description

Sends a single user prompt to the OpenRouter chat completions API.

Usage

query_openrouter(
  prompt,
  model = "openrouter/free",
  temperature = 0,
  top_p = 1,
  max_tokens = 2048L,
  reasoning = TRUE,
  logprobs = FALSE,
  top_logprobs = NULL,
  auto_router_cost_tier = NULL,
  session_id = NULL,
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = Sys.getenv("OPENROUTER_API_URL", unset =
    "https://openrouter.ai/api/v1/chat/completions"),
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
reasoning

Whether to enable reasoning mode.
logprobs

Whether to return log probabilities for output tokens.
top_logprobs

Optional number of most likely tokens to return at each position. Must be an integer from 0 through 20 and requires logprobs = TRUE.
auto_router_cost_tier

Optional OpenRouter Auto Router cost tier: "low", "medium", "high", "xhigh", or "max". Applies when model = "openrouter/auto".
session_id

Optional OpenRouter session identifier for preserving model selection across related requests.
api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string by default, or a parsed JSON list when json_list = TRUE.

Query an OpenRouter Multimodal Model

Description

Sends a single OpenRouter chat-completions request using either simple text content or an explicit multimodal content block list.

Usage

query_openrouter_content(
  content,
  model = "openrouter/free",
  temperature = 0,
  top_p = 1,
  max_tokens = 2048L,
  reasoning = TRUE,
  logprobs = FALSE,
  top_logprobs = NULL,
  auto_router_cost_tier = NULL,
  session_id = NULL,
  modalities = NULL,
  api_key = Sys.getenv("OPENROUTER_API_KEY"),
  url = Sys.getenv("OPENROUTER_API_URL", unset =
    "https://openrouter.ai/api/v1/chat/completions"),
  json_list = FALSE,
  timeout = 120
)

Arguments

content

A non-empty character string or a non-empty list of OpenAI style content blocks.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
reasoning

Whether to enable reasoning mode.
logprobs

Whether to return log probabilities for output tokens.
top_logprobs

Optional number of most likely tokens to return at each position. Must be an integer from 0 through 20 and requires logprobs = TRUE.
auto_router_cost_tier

Optional OpenRouter Auto Router cost tier: "low", "medium", "high", "xhigh", or "max". Applies when model = "openrouter/auto".
session_id

Optional OpenRouter session identifier for preserving model selection across related requests.
modalities

Optional output modalities, for example "text" or c("text", "image").
api_key

OpenRouter API key. Defaults to Sys.getenv("OPENROUTER_API_KEY").
url

OpenRouter chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string by default when the assistant returns text, the first image payload or URL when text is absent but images are returned, or the parsed JSON response when json_list = TRUE.

Query a Qwen Model

Description

Sends a prompt to Alibaba Cloud Model Studio's OpenAI-compatible chat endpoint.

Usage

query_qwen(
  prompt,
  model = "qwen-plus",
  temperature = 0.7,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("DASHSCOPE_API_KEY"),
  url = Sys.getenv("QWEN_API_URL", unset =
    "https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions"),
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

Qwen API key. Defaults to Sys.getenv("DASHSCOPE_API_KEY").
url

Qwen OpenAI-compatible chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

Value

A character string by default, or a parsed JSON list when json_list = TRUE.

See Also

list_qwen_models()

Query a Zhipu Model

Description

Sends a prompt to the Zhipu AI Open Platform's OpenAI-compatible chat endpoint.

Usage

query_zhipu(
  prompt,
  model = "glm-5",
  temperature = 0.7,
  top_p = 1,
  max_tokens = 2048L,
  api_key = Sys.getenv("ZHIPUAI_API_KEY"),
  url = "https://open.bigmodel.cn/api/paas/v4/chat/completions",
  json_list = FALSE,
  timeout = 120
)

Arguments

prompt

A non-empty character string.
model

Model identifier.
temperature

Sampling temperature.
top_p

Nucleus sampling parameter.
max_tokens

Maximum number of output tokens.
api_key

Zhipu AI API key. Defaults to Sys.getenv("ZHIPUAI_API_KEY").
url

Zhipu OpenAI-compatible chat completions endpoint.
json_list

If TRUE, return the parsed JSON response as a list.
timeout

Maximum number of seconds to wait for the HTTP request.

See Also

list_zhipu_models()

Shell Script Helpers

Description

inferencer also ships a small optional shell layer under inst/shell. These scripts are executable command-line helpers that mirror common R wrapper workflows: query_gemini, query_groq, query_openrouter, query_ollama, query_fallback, query_local_llm (a Node.js helper for a local model server), provider-specific ⁠list_*_models⁠ scripts, and render_markdown_terminal.

Details

The shell scripts read API keys and defaults from shell environment variables, typically configured in .zprofile. The variable names are aligned with the R wrappers, such as GEMINI_API_KEY, GROQ_API_KEY, OPENROUTER_API_KEY, and OLLAMA_API_KEY.

The zsh scripts are intentionally minimal and depend on external curl and jq binaries. query_local_llm requires Node.js 18 or later and uses its built-in fetch. Query scripts print the model response text by default and return the full parsed JSON payload when called with --json. Pipe query output to render_markdown_terminal when terminal markdown rendering is desired. query_fallback tries query_gemini, then query_openrouter, then query_groq using each script's default model.

Write Gemini Audio to a File

Description

Decodes base64 audio returned by query_gemini() and writes it as raw PCM or a WAV file.

Usage

write_gemini_audio(
  x,
  path,
  format = c("pcm", "wav"),
  sample_rate = 24000L,
  channels = 1L,
  bits_per_sample = 16L
)

Arguments

x

A base64-encoded audio string, or a parsed JSON response returned by query_gemini(..., json_list = TRUE).
path

Output file path.
format

Output format: "pcm" or "wav".
sample_rate

Sample rate used when writing WAV output.
channels

Number of audio channels used when writing WAV output.
bits_per_sample

Bit depth used when writing WAV output.

Value

Invisibly returns path.