DrPurple27/Project_Ayassek — reverse-engineered prompt

Build me a local AI assistant called Ayassek that runs with one Python entry point and opens a clean dark mode web app on port 2727. I want it to feel like a general brain agent, not just a basic chat box, so it should stream replies live, let me switch between different model providers from the UI, and work especially well with a local Ollama setup, while still supporting OpenAI, vLLM, and NVIDIA NIM if keys or local servers are available.

In the web app, give me normal chat, a model selector, a memory panel, logs, and visible tool activity. The agent should be able to use tools like web search, shell commands, system info, and simple memory store and recall, with a loop where it can plan, act, reflect, and continue. Keep data inside the project folder so it’s portable, and make it reachable from other devices on my network.

Please wire up the env config from the example file and make the default experience just work. Look up current docs online if you need to.