ES version is available. Content is displayed in original English for accuracy.
Hi everyone,
Been working on Otis, an open-source ai agent that gives you one minimal experience across local and hosted open-weight models, privacy-focused by design.
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. Ollama, LM Studio and Nvidia PAIR are supported too.
Excited for everyone to try it and all feedback is welcome!
Discussion (4 Comments)Read Original on HackerNews
I see you are reliant on ollama and llama.cpp (for example), so why should I not just use those directly if I am focused on local first models. (llama.cpp already makes recommendations on what models to run.)
If the argument is then that you are more of an agentic experience, then how are you differentiating yourself from something like the Hermes Agent?
I think more reasons on why this vs. what is already out there would be needed.
Tbh Otis started as a personal TUI agent because I wanted something adapted to how I like to visualize things. Then from there it turned into the unified experience across local and hosted models to make it more accessible and intuitive for anyone. There is some overlap with Hermes but the interface is quite complex for most users.
I don’t believe in sacrificing functionality but at the same time you should not need to understand and learn model runtimes and harnesses etc. to use powerful AI successfully. Ollama and similar tools give users a bunch of models, options and settings to sift through but with Otis the focus is on simplicity, abstracting away the complexity to make getting right to work easy.
I agree though that I need to do a better job of communicating Otis’ differentiators. FWIW I am thinking about additional features to further simplify the experience, like automatic model routing.