openinterpreter: A coding agent optimized for low-cost models.
What problem it solves
Coding-agent workflows can depend on model-specific harnesses, provider settings, and editor integrations. The project addresses these needs with selectable harnesses, provider switching, and compatibility with shared agent protocols and the Codex exec protocol. It also aims to keep instructions, skills, and configuration portable across compatible agent setups.
Key features
Selectable model harnesses — Users can inspect or switch the active Rust-native model harness from the terminal interface and change providers and models as needed.
Source: README
Sandboxed command execution — Commands can run inside native sandboxing across macOS, Linux, and Windows.
Source: README
ACP and Codex compatibility — The agent can run in ACP-compatible editors and clients, while Codex SDK users can select it through a binary override.
Source: README, sdk
Computer Use through a QA skill — A built-in QA skill lets models operate and test interfaces, including web applications and native applications.
Source: README
Portable agent configuration — Shared instructions and skills can live in repository `AGENTS.md` files and `.agents/skills` directories, while product-specific configuration and session state remain local.
Source: README, docs
Quick start
Copied verbatim from the project's README.
curl -fsSL https://www.openinterpreter.com/install | shirm https://www.openinterpreter.com/install.ps1 | iexWho it may suit
Inference: may suit developers who want a terminal coding agent that supports low-cost models, multiple harnesses, and portable agent configuration. It may also suit teams integrating coding agents with ACP-compatible editors or the Codex SDK.
Topics to discuss
Which additional shared agent standards should the project prioritize for portability?
How should users evaluate and compare the available model harnesses?
What workflows would be most useful to cover with the built-in QA and computer-use skills?
Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.
Read the original README https://github.com/openinterpreter/openinterpreter#readme
ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.
Comments
…