gao-ruilin/autorun

★ 101PythonAudience · developerComplexity · 2/5Setup · easy

Mindmap

mindmap
  root((AutoRun))
    What it does
      AI coding assistant
      Terminal chat
      Web UI option
    APIs supported
      OpenAI compatible
      Anthropic compatible
    Usage modes
      Interactive REPL
      Web browser UI
      Pipe mode
    Setup
      Python 3.8+
      Install script
      Config wizard

mindmap root((AutoRun)) What it does AI coding assistant Terminal chat Web UI option APIs supported OpenAI compatible Anthropic compatible Usage modes Interactive REPL Web browser UI Pipe mode Setup Python 3.8+ Install script Config wizard

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Chat with an AI model in your terminal without opening a browser, using any OpenAI or Anthropic-compatible API endpoint

USE CASE 2

Run AutoRun in web UI mode to get a browser-based chat interface on localhost for a more visual coding assistant experience

USE CASE 3

Pipe a code snippet or question to AutoRun in non-interactive mode to get an AI answer as plain text for use in scripts

USE CASE 4

Switch between different AI providers or models mid-session using in-REPL commands without restarting the tool

Tech stack

Python

Getting it running

Difficulty · easy Time to first run · 5min

Run the provided install.bat (Windows) or install.sh (macOS/Linux) to create a venv and register the autorun command.

In plain English

AutoRun is a command-line AI coding assistant written in Python that connects to any AI service compatible with the OpenAI or Anthropic API format. You configure an API endpoint and key, then run it from a terminal to ask coding questions, get explanations, or request help with code in an interactive session. The tool supports three methods for API configuration. A first-run setup wizard walks you through entering the API type, base URL, key, and model name on first launch. The same information can be provided through environment variables set before launching. Or you can configure the connection from inside the running session using short commands such as /api and /model. Configuration is stored in a file in your home directory so it persists between sessions. Once running, AutoRun offers three modes of use. The default is an interactive REPL where you type questions and receive answers in the terminal. A web UI mode starts a local server at a configurable port, accessible in your browser, that provides a graphical chat interface for the same underlying functionality. A pipe mode lets you pass a single question on the command line and receive the answer as plain text, which makes it straightforward to integrate into scripts or pipelines. Installation requires Python 3.8 or newer. The recommended approach is a provided install script (a batch file on Windows, a shell script on macOS and Linux) that creates a virtual environment and registers a global autorun command. Manual installation via pip is also supported. The README includes a bilingual section with the same content in Chinese for Mandarin-speaking users. The README does not mention a license.

Copy-paste prompts

Prompt 1

Using AutoRun with an Anthropic API key, set it up to connect to Claude and explain how to configure the API type and model inside the REPL. Then show how to ask it to review a Python function.

Prompt 2

I want to use AutoRun's pipe mode to get AI explanations of command output in a shell script. Show a complete example pipeline that pipes a Python error message to AutoRun and captures the answer.

Prompt 3

How do I switch AutoRun from an OpenAI API to a self-hosted model running via Ollama? What API type and base URL do I configure?

Prompt 4

Set up AutoRun's web UI mode on a custom port so I can access it from another device on my local network. What flags do I pass and what are the security considerations?

Open on GitHub → Explain another repo

← gao-ruilin on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.