Post
61
I built Otis, a minimal AI agent so you can put those HF models to work!
You can ask it to research a question, work on a document, or fix code in a project. There’s a desktop app (Mac/Linux) and a terminal interface.
On setup it recommends a local model based on the hardware Otis is running on, downloads it and runs it through llama.cpp for you. oMLX, Ollama, LM Studio and Nvidia PAIR are supported too.
Otis can update a document, code and make things, show the results beside your conversation, and save revisions you can return to. The files stay on your computer.
In the attached screenshot, I asked Otis to create a dummy resume for a role using Qwen 3.8 27B 4 bit running on M4 Max 36 GB MBP using oMLX. The work took 6 minutes at 17.8 tok/s.
Excited for everyone to try it and all feedback is welcome!
https://triangllabs.ai/otis
You can ask it to research a question, work on a document, or fix code in a project. There’s a desktop app (Mac/Linux) and a terminal interface.
On setup it recommends a local model based on the hardware Otis is running on, downloads it and runs it through llama.cpp for you. oMLX, Ollama, LM Studio and Nvidia PAIR are supported too.
Otis can update a document, code and make things, show the results beside your conversation, and save revisions you can return to. The files stay on your computer.
In the attached screenshot, I asked Otis to create a dummy resume for a role using Qwen 3.8 27B 4 bit running on M4 Max 36 GB MBP using oMLX. The work took 6 minutes at 17.8 tok/s.
Excited for everyone to try it and all feedback is welcome!
https://triangllabs.ai/otis