FAQ

What is Nexis?

An open-source, lightweight, cross-platform AI-native terminal and developer environment. Terminal, editor, language tooling, source control, and an AI panel in one app. See What is Nexis?.

What is it built with?

Tauri 2, Rust, and React 19. The native PTY and heavy lifting are in Rust; the UI renders in the system webview, which is why the app is under 10 MB.

Is it free and open source?

Yes — Nexis is released under the Apache-2.0 license. It builds on the open-source terax-ai project.

Does Nexis send my data anywhere?

No. There is zero telemetry. Your code stays on your machine, and API keys are stored in the OS keychain, never on disk. If you use a cloud AI provider, only what you send to that provider goes to it — and the AI context inspector shows you exactly what that is.

Can I use it without any AI provider / API key?

Yes. Point Nexis at a local LM Studio, MLX, or Ollama endpoint and it runs fully offline with no key. See AI providers.

Which AI providers are supported?

12+, including OpenAI, Anthropic, Google, Groq, xAI, Cerebras, DeepSeek, Mistral, OpenRouter, Hugging Face, any OpenAI-compatible endpoint, and the local runners above. Full list on AI providers.

Will the AI edit or run things without asking?

No. Every file, shell, search, and plan tool requires explicit approval, and AI-proposed edits are shown as per-hunk diffs you approve or reject individually.

What platforms are supported?

macOS, Linux (.deb / .rpm / AppImage / AUR), and Windows (NSIS / MSI), with first-class WSL. See Installation.

What is the ML Lab?

A panel for training small models on your own data, powered by a local engine — either the Python nexis-ml or the single-binary Rust nexis-ml-rs. See ML Suite.

How do I report a bug or request a feature?

Open an issue on the Nexis GitHub repository.