SuperAPI AI Directory

gpt-runner

GPT-Runner is a tool for managing AI presets and conversing with code files, designed to improve development efficiency for individuals and teams. It offers three ways to use it: CLI, web version, and VSCode extension. It supports selecting project files so AI can answer based on the latest content, and saves AI role presets as version-controllable .gpt.md files. The tool supports custom AI parameters, third-party LLMs, local data storage, and multiple languages.

Open website

About this tool

GPT-Runner is an open-source tool for managing AI presets and conversing with code files to improve development efficiency. It offers three usage forms: CLI, web version, and VSCode extension.

Main uses:

  • Converse with code files: no need to manually copy and paste multiple files' code into the ChatGPT window; simply select project files from the file tree, and the AI will provide answers based on the latest content of these files.
  • Manage project AI presets: each xxx.gpt.md file represents an AI role preset, is easy to read and modify, and can be version-controlled, making it convenient for team members to share and enhance AI presets.

Main capabilities:

  • Converse with code files: select files or folders and have real-time conversations with AI.
  • Powerful CLI and IDE integration: implement efficient AI workflows in various IDEs.
  • AI preset manager: manage AI presets like a local Storybook.
  • Custom AI parameters: flexibly control the configuration of AI models.
  • Support third-party LLMs: high compatibility and adaptability.
  • Privacy first: local data storage protects privacy.
  • Multilingual support.

Usage:

  • Requires an OpenAI Key or Anthropic Key.
  • CLI method: requires NodeJS >= 16.15.0, run the command to download the package (specific command not provided), then visit in the browser http://localhost:3003 to view the web interface.
  • VSCode extension method: requires VSCode >= 1.78.0, install GPT-Runner VSCode Extension from the VSCode Marketplace.
  • Configuration files: involves the gptr.config.json configuration file, xxx.gpt.md AI preset files, and the special .gpt-runner directory; detailed documentation can be found in the project introduction.

Applicable scenarios:

  • Development teams need to share and version-control AI presets.
  • Developers want to converse directly with code files to propose features or fix bugs.
  • Need to integrate AI workflows in multiple IDEs.

Limitations:

  • Requires your own OpenAI or Anthropic API Key.
  • The CLI method requires a NodeJS environment, and the VSCode extension requires a specific version.
  • Some features (such as the Jetbrains plugin, chat history import/export, AI preset store, template interpolation, and Electron client) are listed in the documentation as to-dos or future plans and may not currently be available.
  • Specific installation commands, configuration details, and pricing information are not yet provided.