OpenShell is NVIDIA's open-source runtime for fleets of autonomous AI agents. Written in Rust and licensed under Apache 2.0, it gives agents a sandboxed place to execute with isolation primitives and an extension surface, so a team can let agents act on real systems without handing over the host. It installs from PyPI as the openshell package and ships published documentation and a 0.1.x upgrade guide.
Be the first to rate OpenShell. Your feedback helps other users decide.
Agents execute inside an isolated runtime instead of directly on the developer machine.
The runtime is built for fleets of agents rather than one long-running conversation.
Version 0.1.x adds new isolation primitives and an expanded extension surface with a documented upgrade path.
Similar tools in the same category — compare side-by-side
MIKE (Integrated Kanban Environment) is an open-source LLM harness for software development work. It...
Corgi is a terminal multiplexer for humans and agents. It keeps watch over every agent you have runn...
Craig is a keyboard-driven terminal workspace that runs Claude Code, Codex and Cursor in parallel, e...
Guides and comparisons about code tools
Eight new tools that decide where a coding agent runs, how many run at once, and how you keep working in the same file it is editing.
Twenty-five coding tools sorted by what they do: autocomplete, codebase comprehension, testing, or self-hostable models, from CodeGeeX to Codestral.
Coding agents increased output but not shipped software. The gap is verification, context and drift. Here is what converts agent output into working systems.
Agents need guardrails. Blocklists and judge models fail because they cannot see data flow. Deterministic engines track labels across tool calls instead.
How development environments built around delegation differ from autocomplete-era assistants, and what Windsurf and Google Antigravity each do well.
