LLM Studio

Make room for AI that runs on your device. LLM Studio brings local language models, rich conversations and a code workspace to Android.

YOUR MODELS, YOUR CHOICE

Browse the model catalog or open a compatible GGUF file from your phone. Keep downloaded models in a folder you choose. Model availability, size and hardware requirements vary.

A CHAT BUILT FOR IDEAS

Write, learn and explore with local models. Read Markdown, tables, formulas and highlighted code. Organize conversations with folders, search and pins. Attach images and open them in a full-screen viewer; image understanding requires a compatible vision model and its companion files.

FROM CODE TO PREVIEW

Run supported JavaScript and Python snippets in the built-in offline environment. Open supported HTML, CSS and JavaScript projects as interactive web previews directly from a response. This is a browser-based sandbox, not a server or a full desktop development environment. Python availability depends on your Android WebView.

CONTROL YOUR LOCAL RUNTIME

Choose CPU or supported GPU and NPU options, or use automatic backend selection. Availability and performance depend on the device, drivers, model and available memory. CPU remains the compatibility option.

ON-DEVICE CONVERSATIONS

Local inference does not require a LLM Studio account or a cloud AI service. Chats and attachments stay on your device. Internet access is needed for model downloads and optional AI-response reports. A report sends the selected reason and optional note; the response text is included only if you choose to attach it. Version 1.0 is free, with no ads or in-app purchases.

The interface supports English, Ukrainian, Russian and nine other languages.

Before you start: download or import a compatible model. Models can occupy several gigabytes, and larger models require more RAM. Responses may be inaccurate; verify important information. This is a general-purpose AI tool, not a medical, financial or legal advisory service.

Privacy policy: https://vans.wtf/llm-studio/privacy.html

Contact: evhenii.svistunov@gmail.com