Tools for
Conviviality

Home / Coding agents / Open Interpreter

Open Interpreter

A terminal coding agent tuned to get strong results from cheaper open models, with commands run in a sandbox.

Instead of Claude Code, Codex

M · ToolOpen sourceFreeBring your own model
github.com/openinterpreter/openinterpreter
Screenshot of the Open Interpreter website

Screenshot from github.com/openinterpreter/openinterpreter, taken 2026-10-30, shown to help you recognise the tool. It belongs to its maker. Ask us to remove it.

Open Interpreter began as a Python project that let a model run code on your computer, and it has since been rebuilt from scratch, a good lesson in how fast this field moves. The current repository describes it as a coding agent for open models like Kimi K3 and GLM 5.3, a fork of OpenAI's Codex that imitates other agents' harnesses to get better results from low-cost models. You switch harnesses with /harness and models with /model.

Open Interpreter scores a 5 on lock-in, a 4 on footprint, self-direction, power and delivery and a 3 on longevity.

The code is Apache-2.0 and runs on Mac, Linux and Windows, with commands executed in a native sandbox on all three and state kept under ~/.openinterpreter on your machine. The page names no people or company, and says the original Python version lives on only as a community fork, so Longevity is a 3 for a project that has changed shape. It works with hosted providers that need credentials and with any OpenAI-compatible provider, and the page does not say whether the models themselves run on your machine.

    What stands out

  • A terminal coding agent built to get good results from cheaper open models
  • Apache-2.0, a fork of Codex, with commands run in a sandbox
  • The original Python project survives only as a community fork
Visit Open Interpreter
Do you make Open Interpreter? Get the badge for your site

This badge reports a dated review. It is not a certification and it costs nothing. Link it to this page and leave the numbers alone: the image always shows the current score.

For dark backgrounds, use open-interpreter-dark.svg in place of open-interpreter-light.svg.

For comparison, we also reviewed GitHub Copilot (15/30) and OpenAI Codex (20/30) and Claude Code (17/30) as a benchmark.

Other tools for “code with an ai agent i control”