raix connects RStudio to any AI model. Describe what you want in English — raix writes the R code, executes it, and shows results. Chat, debug, generate notebooks, run Python, compile C++, or launch a full coding dashboard.

remotes::install_github("twomathematicians-code/raix")
library(raix)
raix_dashboard()    # self-configuring — auto-detects your AI

Installation help →  |  Release notes →

Architecture

flowchart TB
    U[User Input] --> SETUP

    subgraph SETUP[Setup]
        A1[raix_setup · raix_configure]
        A2[raix_small_mode · raix_config]
    end

    SETUP --> CORE

    subgraph CORE[AI Backend]
        B1[OpenAI / Groq / Mistral / DeepSeek]
        B2[Ollama / LM Studio / vLLM]
        B3[Claude / Custom endpoints]
    end

    CORE --> APPS

    subgraph APPS[Applications]
        C1[Chat & Dashboard]
        C2[Code Gen & Debug]
        C3[Developer Agent]
        C4[Compute Engine]
        C5[Cross-Language]
        C6[Data & Project]
    end

    APPS --> OUT
    OUT[Scripts · Notebooks · Plots · Results]

Loading

Commands by Workflow

Get Started

raix_setup()          # auto-detect & configure — one click
raix_dashboard()      # full coding workspace in RStudio Viewer
raix_info()           # show current config
raix_config()         # get config as list (for scripts)

Chat & Generate Code

raix_chat()           # interactive AI chat in console
raix_gui()            # chat window in RStudio Viewer
raix_send("prompt")   # one message → one response
raix_explain("code")  # AI explains R code
raix_generate("task") # English → R code
raix_debug()          # diagnose last error
raix_document("fn")   # generate roxygen2 docs

Developer Agent

raix_solve("problem")        # complete solution from description
raix_script("task", "out.R") # generate .R file
raix_notebook("task", "out.Rmd") # generate .Rmd notebook
raix_test(my_function)       # AI writes testthat tests
raix_refactor(code)          # AI suggests improvements
raix_project(".")            # scan project for AI context
raix_package("task")         # find best R package
raix_read("file.R")          # read file with AI summary
raix_write("desc", "out.R")  # AI content → file

Compute & Cross-Language

raix_terminal("cmd")         # shell commands + AI analysis
raix_python("task")          # generate & run Python from R
raix_compile("task")         # C++ with Rcpp, compile, call from R
raix_pipeline(steps)         # multi-step cross-language workflow
raix_translate(code, to="python") # R ↔ Python
raix_sql("top 10 customers") # English → SQL
raix_simulate("Monte Carlo") # AI simulation engine
raix_web("https://...")      # fetch & summarize web pages

Performance & Data

raix_benchmark(expr)         # time code, AI suggests optimizations
raix_parallel("task")        # AI rewrites for multi-core
raix_analyze(mtcars)         # AI-guided data analysis
raix_search("topic")         # search CRAN packages
raix_diagnose("script.R")    # scan for issues & anti-patterns
raix_google("query")         # Google search + AI summary
raix_sysinfo()               # system info for AI context
raix_history()               # view/search chat history

Developer Agent Flow

flowchart LR
    A[English Description] --> B[raix_solve]
    B --> C[Plan & Select Packages]
    C --> D[Generate R Code]
    D --> E{Output}
    E --> F[.R Script]
    E --> G[.Rmd Notebook]
    E --> H[Execute & Return]
    F --> I[Open in RStudio]
    G --> I
    H --> I

Loading

Supported Models

Local (free)	Cloud (API key)
Ollama · LM Studio · vLLM	OpenAI · Claude · Groq · Mistral
	DeepSeek · Together AI · Perplexity · OpenRouter

Small models? raix auto-detects 7B-9B models and optimizes prompts. raix_small_mode(TRUE) for manual control.

License

MIT © Mahesh Solanki · SOP.md · citation("raix")