Open Interpreter SOP

Appearance

Install ​

Open Interpreter ships as a pre-built native binary installed via a one-line script. No Python or Node runtime is required for the Rust version.

macOS and Linux ​

bash
curl -fsSL https://www.openinterpreter.com/install | sh

Windows (PowerShell) ​

powershell
irm https://www.openinterpreter.com/install.ps1 | iex

After the script completes, the interpreter binary (also aliased as i) is available in your PATH.

What gets installed ​

  • The interpreter binary placed in a system-appropriate location (e.g., ~/.local/bin on Linux/macOS)
  • Config and session state directory created at ~/.openinterpreter

Setting up a model provider ​

Open Interpreter needs a model API key to function. Set the relevant environment variable before starting a session:

bash
# Example: OpenAI
export OPENAI_API_KEY=sk-...

# Example: Anthropic
export ANTHROPIC_API_KEY=sk-ant-...

# Example: Deepseek
export DEEPSEEK_API_KEY=...

See the provider docs for the full list of supported providers.

No downloadable template or boilerplate ​

Open Interpreter does not ship a project template or starter repo. Configuration is done via flags, the /model and /harness TUI commands, and an optional AGENTS.md file in your project root.