TT Studio

Run Local AI on your Tenstorrent hardware.

AI model management in an easy-to-use GUI. Privately run LLMs, voice agents, image/video generation and more without paying for tokens.

View on Github ↗
TT-Studio chat interface showing the start of a conversation with the Qwen3-32B model

Up and running on your TT-QuietBox® 2 in one command.

Copy 
$ tt-studio

Have a Tenstorrent card? Follow the setup instructions in the GitHub repo.

One stack. Every modality.

Everything you need to deploy, chat with, and build on AI models, running privately on your own Tenstorrent hardware.

Chat with LLMs
Deploy a model and start talking. Llama and more, served on your cards.
Voice agent
Speak and listen, with wake-word and voice-activity detection built in.
Media generation
Generate images and video, entirely on-device.
RAG
Query your documents with retrieval augmented generation.
Remote endpoints
No Tenstorrent card? Use models running on cards elsewhere.
One-command setup
tt-studio handles Docker, your env, and the whole stack.

Other resources.

More open-source tools from Tenstorrent.

TT Developer Toolkit ↗
Interactive AI lessons, hardware monitoring, and production inference templates, all inside your editor.
tt-toplike ↗
A psychedelic, ASCII-native terminal monitor for Tenstorrent Blackhole and Wormhole hardware.
tt-local-generator ↗
Local AI image and video generation for builders, running entirely on your own hardware.