urazerkoc3

yoo guyz, sorry for waiting but im trying to make it usable for everyone.

News:

• I improved system prompt
• Changed skill system and i write it from scratch again
• Basic Skills turned into .dll file so not need .pyd files anymore
• Added TTS, i used edge-tts for it and it’s works verry well

Future Plans:

• I’ll add browser skills and more
• About long term and short term memories, i’ll add these stuffs as a tool
• STT (Speech to Text) actually it’s unnecessary so this plan will be last one i think
• And a good ui for make it more easy to use

i’ve added executer and improved system prompt

i tried to make local system more effective but it didn`t work and i’ve added Json catcher.

I’ve just started coding main project, for now i’ve added gemini api as a lib and groq api as a lib and some system prompt improvements

in video, you can see grok api responses with meta-llama/llama-4-scout-17b-16e-instruct LLM

I Turned local to api system because of Local Ai is consume to much sources and its not worth to run it in local.

i’ll use groq for testings but also i’ll integrate gemini api

Oryxis, a Jarvis Like Self-Learning Desktop Ai Assistant

its runs in local and uses llama-cpp-2.rs for running local ai.