Building AI Assistance Model System

[PRICE IS A PLACEHOLDER] Build a Unified Local AI Assistant Platform I am looking for someone to build a complete desktop application that allows me to run multiple specialized AI models through one clean, easy-to-use interface. The goal is for the entire system to behave like separate AI with 3D AI faces and voices as multiple specialists working together behind the scenes. Each AI should have its own icon that I can click to speak directly with that specialist. Additionally, there should be one Master Launch icon that starts the entire system at once and allows all of the AI models to work together collaboratively. The finished application should require as little manual setup as possible after installation. AI Models The system will use the following models: Qwen3-Coder Fin-R1 Hermes-4 FLUX.1-dev Magnum V4 Qwen3-235B Kimi K2.5 Large models may run through hourly cloud GPU services while lighter models may run locally if appropriate. Unified AI System Although each model has its own specialty, I do not want them functioning like completely separate applications. Instead, I want one interconnected system where: Every model can communicate with one another. The models can pass tasks between themselves. One model can request assistance from another. Shared memory can be accessed when appropriate. Specialized memories can remain private to individual models when needed. For example: The writing model can ask the coding model for code. The research model can provide information to the writing model. The finance model can analyze financial data generated elsewhere. The image model can generate artwork requested during conversations. Everything should feel like one AI made up of multiple specialists. Interface The application should have a clean desktop interface. I would like: One Master Launch icon that starts the complete system. Individual icons for every AI model. Ability to switch between specialists instantly. Status indicators showing whether each model is online. Simple settings menu. Characters Each AI should have: Its own avatar or face. Its own voice. Separate personality and system prompt. NVIDIA Audio2Face lip syncing if possible. Chatterbox TTS voices. Speech recognition for talking naturally with the AI. Voice Features Required features: Voice recognition (speech-to-text) Natural text-to-speech Push-to-talk Optional wake word if practical Real-time conversations Ability to interrupt speech naturally Memory Use: PostgreSQL ChromaDB The system should support: Long-term memory Conversation history Shared memories Individual memories Searchable memories Controller Python is preferred for the backend. Responsibilities include: Starting every service Routing prompts Managing memory Coordinating the AI models Cloud GPU communication WebSocket communication Voice processing Launching and monitoring the entire system Hardware Primary computer: ASUS ROG Ally Storage: External 1–4 TB SSD The application should be configured to load models from the external SSD whenever possible. Cloud GPU Support The system should support hourly cloud GPU services for the larger models. The local interface should remain responsive while the heavy AI models run remotely. If possible, cloud providers should be easy to change later without rebuilding the application. Installation Requirements One of the most important deliverables is a complete installation process. I would like: A step-by-step installation guide written for someone with limited programming experience. Instructions for downloading every required model. Instructions for organizing the models on the external SSD. Instructions for installing every dependency. Instructions for configuring cloud GPU services. Instructions for setting up PostgreSQL. Instructions for setting up ChromaDB. Instructions for installing Python and required packages. Instructions for configuring WebSockets. Instructions for connecting Chatterbox TTS. Instructions for connecting NVIDIA Audio2Face. Instructions for changing voices later. Instructions for replacing avatars later. Instructions for replacing AI models later. Instructions for updating the system in the future. The goal is that I can completely rebuild the system later using only the documentation provided. Ease of Use The finished application should feel like a polished desktop application rather than a collection of scripts. Ideally I should be able to: Double-click one icon. Wait for everything to load automatically. Begin talking to the AI immediately. No manual terminal commands should be required during normal use. Deliverables Complete source code Full documentation Installer or setup package if possible Configuration files Step-by-step installation guide Model installation guide Cloud GPU setup guide External SSD setup guide Future update instructions Instructions for adding new AI models Instructions for replacing existing models Instructions for changing personalities Instructions for changing voices Instructions for replacing avatars Please tell me how you would build the project, which technologies you recommend, your estimated timeline, and examples of similar work if available.