LM Studio Releases Bionic Coding Agent for Linux

LM Studio has expanded its software lineup by releasing the Linux version of Bionic, its dedicated agentic application. Previously restricted to macOS and Windows, the tool is now available for developers running Ubuntu, Fedora, or Arch. Bionic operates as a standalone application distinct from the classic LM Studio, which continues to serve as a local model browser and OpenAI-compatible inference server on localhost:1234.

The application is designed specifically for complex tasks, research, and coding workflows using open-source models. Powered by the same runtime engine as LM Studio, which utilizes MLX and llama.cpp, Bionic runs locally to support GGUF formats and on-device voice transcription. It features two primary workspaces: a Code project that integrates with local folders for repository-aware editing with inline diffs, and a Work project for managing documents, PDFs, slides, and spreadsheets. Every edit is autosaved, allowing users to hand off tasks to the agent seamlessly.

While Bionic prioritizes local execution, it also connects to the LM Studio Secure Cloud for demanding tasks. This paid, metered tier offers zero data retention guarantees and runs frontier open-weight models like GLM 5.2, Kimi K3, and DeepSeek V4 Pro that exceed local hardware limits. Otherwise, local inference remains entirely free, costing nothing beyond the user's existing hardware and electricity.

This release addresses a major request from developers who rely on Linux-based GPU workstations and home labs. For compliance-sensitive teams that cannot send proprietary code to external APIs, Bionic offers a secure, private alternative. By pairing the app with LM Link over a Tailscale tunnel, users can route tasks from a laptop to a more powerful remote tower. Developers can install the application for free via the official website or by running a curl command in their terminal.