UE 5.5.4 Project with LLM Integration

I need to build a UE 5.5.4 project that utilizes either: 1- A local LLM (Gemma 4 E2B or Gemma 4 E4B) prompting it in a multi-model fashion 2- An online service (OpenAI's GPT models, Anthropic's Claude, Google's Gemini, xAI's Grok, DeepSeek) I can switch between Local and Online as a user and for each mode I can choose the needed model. The project has one scene with (Ultra Dynamic Sky) and a single slider widget that control the time of day and a selector to choose between its existing weather presets (clear, rainy, foggy, snowy, etc...). This manual part is already done. I need to add a prompt box at the bottom of the screen that allows me to prompt the LLM (Local or Online) with multi-modal input: - Text (In any language) - Image file (Picked from File Browser) - Viewport screenshot of the current 3D scene (without the UI) - Audio file (Picked from File Browser) - Mic input (Spoken in any language) and the LLM will respond with: - Translate the text from whatever language it was to English (for internal processing) - Changing the time of day slider and weather selector to the requested value (Example: If I tell it set the time to sunset or set it to 18:34 or set the weather to be cloudy, it will do it in both cases) - Transcribe the audio file to text (and then translate that to English text) - Analyze and describe an image (either the image file or the viewport capture). For example if the image is of a rainy environment or of a sunset (or both) then I can tell it "Match the time of day and weather scene in the image" and the LLM will execute the function to change the time of day or weather. - If the user asked a question, it will answer based on a documentation website (GitBook) so its knowledge comes from that specific website. Effectively acting as a Help/Support Bot. The URL of the text file to be fed to either the local or remote LLM for knowledge is (https://creadev.gitbook.io/creadev-docs/llms-full.txt) Note: - If Local LLM mode is selected, I need to be able to switch between Gemma 4 E2B and Gemma 4 E4B on the UI so I can experiment with both. - if Online LLM mode is selected, I need to be able to switch between (OpenAI's GPT models, Anthropic's Claude, Google's Gemini, xAI's Grok, DeepSeek) and provide an API key for each of them. - All of this should be working in a standalone shipping build on Windows, Mac, Linux, iOS, Android. - There should NOT be a requirement to have a separate Ollama Server running, it should be fully self contained (the .gguf file is embedded in the executable) - Must work with Unreal Engine 5.5.4 - Use Blueprint, not C++ - Please submit your estimated time and cost with your proposal.