LM Studio Bionic
LM Studio Bionic is a natively local AI agent for open models that handles coding, documents, slides, and PDFs on your own machine, with optional frontier open cloud models like GLM and Kimi run under zero data retention.
At a Glance
Get started with LM Studio Bionic at no cost with Use local LLMs and voice transcription on your own machine and Local model running via llama.cpp and MLX.
Engagement
Available On
Alternatives
Listed Jul 2026
About LM Studio Bionic
LM Studio Bionic is an AI agent for open models that runs natively on your own computer. Built by Element Labs, the makers of LM Studio, it can write and edit documents, build slides and PDFs, and take on full coding tasks while keeping data on your machine. For heavier work it can also call frontier open models like GLM and Kimi through LM Studio Secure Cloud with zero data retention. Bionic is a separate app from the original LM Studio and launched in July 2026 for Mac and Windows.
What It Is
Bionic is LM Studio's agentic app, positioned by the vendor as "the AI agent for open models." Where the original LM Studio app centers on chatting with and managing local models, Bionic wraps those same open models in an agent that can use tools, edit files, and complete multi-step tasks. The homepage frames it as being "built for creativity, work, and code," and the launch blog describes it as the agent "for getting real work done with open models." It organizes work into Work Projects for research, writing, and document tasks and Code Projects that point at a local folder, and it lets you run several sessions in parallel across projects.
Working With Documents and Code
For everyday work, Bionic can create and edit documents, slides, and PDFs, and according to the docs every change is saved automatically so you can iterate without manually managing versions. Document processing runs in a sandboxed environment. On the coding side, the vendor says Bionic behaves like a full coding agent, with file, search, Git, and shell tools, the ability to inspect a local codebase, inline diffs for reviewing changes, and automatic checkpoints so you can roll back safely. A built-in web search tool can pull in external context when a task needs it.
Local Voice Transcription
Bionic includes what the vendor calls state-of-the-art local voice transcription so you can talk through ideas, prompts, and edits and have speech turned into text in real time. The launch materials attribute the transcription to Mistral AI's Voxtral model, describe it as multilingual, and note it works as a voice keyboard across any app. The company states that speech and audio are processed locally and never leave your device.
Local and Cloud Open Models
Bionic can download and run open LLMs directly inside the app, powered by the LM Studio runtime with MLX and llama.cpp under the hood, so simple chats and agentic tasks can run entirely offline. For the most demanding coding, reasoning, tool-calling, and long-context work, the vendor offers frontier open models such as GLM and Kimi through LM Studio Secure Cloud. LM Link lets you connect to models running on other devices, so a session can draw on local, remote, or cloud models depending on the task.
Privacy and Data Handling
The company describes privacy as central to the LM Studio ethos. It says its cloud inference servers are US-based and operate with Zero Data Retention by default, meaning requests are processed transiently and not retained after completion, and it states it does not train on user data. Combined with the local-first design, the vendor's pitch is that you can decide on a per-session basis how much runs on your own hardware versus in the cloud.
Community Discussions
Be the first to start a conversation about LM Studio Bionic
Share your experience with LM Studio Bionic, ask questions, or help others learn from your insights.
Pricing
Free
Get started with LM Studio Bionic at no cost with Use local LLMs and voice transcription on your own machine and Local model running via llama.cpp and MLX.
- Use local LLMs and voice transcription on your own machine
- Local model running via llama.cpp and MLX
- Offline voice transcription
- ZDR web search tool (limits apply)
- LM Link for up to 5 devices
Pay as You Go
Purchase cloud credits to run frontier open models on LM Studio Secure Cloud, US-based with Zero Data Retention by default.
- Frontier open cloud models (GLM 5.2, Kimi K2.6, Kimi Code K2.7, more coming)
- US-based inference with Zero Data Retention by default
- Buy credits as needed
Bionic Pass
Subscription plan announced by the vendor; pricing and plan details listed as coming soon.
- Details announced as coming soon
Capabilities
Key Features
- AI agent for open models, built for work and code
- Create and edit documents, slides, and PDFs with automatic saving
- Agentic coding with file, search, Git, and shell tools
- Local codebase inspection with inline diffs
- Automatic checkpoints for safe rollback
- State-of-the-art local voice transcription (Voxtral by Mistral AI)
- Multilingual voice keyboard usable across any app
- Download and run local open LLMs via MLX and llama.cpp
- Frontier open cloud models (GLM, Kimi) via LM Studio Secure Cloud
- Zero Data Retention on cloud inference by default
- Built-in web search tool for external context
- Work Projects and Code Projects
- Multiple parallel sessions across projects
- LM Link to use models running on remote devices
- Sandboxed document processing
