Open Interpreter
An open-source terminal coding agent for macOS, Linux and Windows. Connect local models through Ollama or LM Studio, or use hosted providers.
Open Interpreter is a terminal coding assistant built around open-weight models, with model-specific behavior for Kimi, Qwen and DeepSeek. It's a fork of OpenAI's Codex for developers who want to choose their model provider while keeping a familiar agent interface. The project uses the Apache 2.0 license.
The agent reads project files, edits code and runs commands on macOS, Linux and Windows. Its OS-level sandbox controls file access and network use, with approval requests when an action needs broader permissions. A built-in QA skill can test web interfaces in a real browser through agent-browser and native apps through trycua.
Local models connect through Ollama or LM Studio and stay on your machine. Cloud options include OpenAI, Amazon Bedrock, Anthropic and OpenRouter, alongside providers for Kimi, Qwen and DeepSeek. With your own API key, requests go directly to the provider. Requests using its hosted model service pass through Open Interpreter's servers and are logged for up to 30 days.
Existing AGENTS.md instructions, shared skills and MCP servers carry over. ACP support lets the agent work in compatible editors, and Codex SDK compatibility lets developers retain existing integrations. Separate coding sessions can resume or fork earlier work.
