vanilla.

the local interface for all flavors of AI.

every flavor of model.

Local through Ollama and LM Studio. Cloud through OpenAI, Anthropic, HuggingFace, and Gemini. Coding CLIs like Aider, Goose, and OpenCode. All unified in one interface.

built for the way you actually work.

Offline voice dictation

Speak instead of type with the powerful Whisper engine that accelerates AI workflows by letting you speak directly to a model.

HuggingFace installer

Browse GGUF models on HuggingFace and import them straight into Ollama without touching the command line.

Files that models can read

Upload documents and images with built-in OCR and image understanding, so any model can work with what you see.

Real-time streaming

Responses stream token by token from every provider, local or cloud, so you’re never staring at a spinner.

Smart auto-naming

Chats name themselves as you go. Add custom system prompts per conversation and export everything when you need it.

Compare side by side

Run the same prompt across two different models at once and watch both stream in real time.

pick your theme.

15 different ice cream-themed color schemes, adjustable text sizes, and a wonderfully customizable UI make Vanilla an easy choice.

and make your own

selected flavor Strawberry

your data never leaves your machine.

Everything runs locally by default. Cloud providers are strictly opt-in. You maintain full control over your AI workflow, without sacrificing privacy or flexibility.

get a scoop.

100% free of charge. runs on your machine.
tested on macOS and Linux.