Quickstart
Open Interpreter is a terminal coding agent built from the Codex CLI surface and adapted for provider-agnostic local use. It can inspect your repository, edit files, run commands, review diffs, and resume work later.
curl -fsSL https://openinterpreter.com/install | shOn Windows, run:
irm https://openinterpreter.com/install.ps1 | iexcd my-project
interpreterThe first run walks you through provider setup. You can sign in with ChatGPT, use an API key, connect a local model through Ollama or LM Studio, or configure another compatible provider.
You can change providers later with /model.
Type a concrete request:
add a /health endpoint that returns the build shaOpen Interpreter reads the project, proposes work, edits files, and runs commands through the active sandbox and approval settings.
In the default local workflow, Open Interpreter can work inside the current
workspace and asks before actions that need more access. Use /permissions
to inspect or change this during a session.
interpreter resume --lastResume keeps the prior conversation, context, and working directory.
Common first commands
| Task | Command |
|---|---|
| Start the TUI | interpreter |
| Start with a prompt | interpreter "explain this repo" |
| Run once without the TUI | interpreter exec "summarize the current diff" |
| Continue the last session | interpreter resume --last |
| Review current changes | /review or interpreter exec review --uncommitted |
| Choose model/provider | /model |
| Change permissions | /permissions |
Next pages
Composer, shortcuts, images, prompts, approvals, review, and background work.
Defaults, profiles, providers, model settings, feature flags, and project config.
How local command execution is constrained and when Open Interpreter asks.
Durable project instructions for build commands, conventions, and cautions.