Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
Commits on main
Studio: dock the floating Live monitor beside Run settings (#11699)
80432b8 Wasim Yousef Said committed 1 month ago
Studio: example prompt per image workflow, remember the last prompt (#11775)
9a8f246 Michael Han committed 1 month ago
Studio: Make `unsloth start --reasoning on/off` take effect (#11718)
0b5b444 Nilay committed 1 month ago
Studio: Keep Codex reasoning between tool calls on /v1/responses (#11726)
346df2b Nilay committed 1 month ago
Studio: fix broken Chinese, Japanese and emoji text in gpt-oss replies (#11720)
7501b31 Nilay committed 1 month ago
Layout contract: tie the Train rail clamp to its scroller's padding (#11773)
4f70162 Daniel Han committed 1 month ago
Studio: stop the max speed tier recompiling on every new prompt length (#11731)
abb11de Daniel Han committed 1 month ago
fix(studio): support preserve thinking for llama.cpp connections (#11706)
4fabb25 Lee Jackson committed 1 month ago
Studio: Show token usage and cache stats on connected provider chats (#11717)
06be728 Nilay committed 1 month ago
Studio: pin, reorder and add to project for Audio history (#11774)
b5100c1 Michael Han committed 1 month ago
Studio: make the HunyuanImage-2.1 denoiser step capturable as a CUDA graph (#11753)
b621579 Daniel Han committed 1 month ago
Studio: let an explicit FBCache request engage on Qwen-Image-2.1 and other prefix-KV models (#11713)
2be14e0 Daniel Han committed 1 month ago
Studio: keep Qwen-Image-2.1 int8 / fp8 compiling on torch 2.12 (CantSplit) (#11677)
69c154d Daniel Han committed 1 month ago
Export full fine-tunes in 16-bit from the CLI by default (#11716)
16f636a Nilay committed 1 month ago
MCP HTTP integration test: let the server bind its own port (#11777)
fb9dc1e Daniel Han committed 1 month ago
studiobench: do not fail UI parity on repetitions that swap two renderings (#11772)
9815e76 Daniel Han committed 1 month ago
Studio: follow-ups for the media gallery, hub link and rail divider (#11765)
168961d Michael Han committed 1 month ago
Studio: stop clipping the model description's descenders; follow the resizable rail in the layout contracts (#11769)
29fc8a9 Daniel Han committed 1 month ago
Account matrix: cover the gallery move and add-to-project routes (#11767)
c90a5ca Daniel Han committed 1 month ago
Update idempotency: count whisper.cpp only when the install has one (#11762)
99dd0f9 Daniel Han committed 1 month ago
Data settings check: pin the legacy store outcome in the failed-delete check (#11763)
f23d84b Daniel Han committed 1 month ago
Studio installer: do not run Apple's git shim to ask whether git works (#11759)
0fc6f17 Daniel Han committed 1 month ago
Studio: make the Images, Video and Audio settings rail resizable (#11760)
298df4f Michael Han committed 1 month ago
Studio: drag to reorder, add to project and quick download for the image and video galleries (#11758)
83db6e0 Michael Han committed 1 month ago
Studio: keep the model name and quant whole in the Images and Video header (#11757)
0ee01af Michael Han committed 1 month ago
Studio: add Search Hub to the Images, Video and Audio model pickers (#11756)
4f3b104 Michael Han committed 1 month ago
Clean-machine trace leg: let git fetch the pinned git+ requirements, nothing else (#11751)
e658273 Daniel Han committed 1 month ago
AMD iGPU routing tests: clear the visibility masks for every test (#11752)
ba18ca7 Daniel Han committed 1 month ago
Chat UI driver: wait for the Recents thread to load, and let the check fail (#11749)
b429f4f Daniel Han committed 1 month ago