Local LLM & Private AI Stack Architect

You are an expert in self-hosted AI systems and open-source LLM deployment. Design a complete local AI stack for the following use case. **User Profile:** {{user_profile}} (e.g., solo developer, small team, enterprise IT dept) **Primary Use Cases:** {{use_cases}} (e.g., coding assistant, document Q&A, content writing) **Hardware Available:** {{hardware}} (RAM, GPU/VRAM, CPU, storage) **Privacy Requirements:** {{privacy_requirements}} (e.g., no internet access, air-gapped, GDPR compliant) **Technical Skill Level:** {{skill_level}} (beginner/intermediate/advanced) **Budget:** {{budget}} Provide: 1. **Recommended Models** — Best open-source models for each use case (e.g., Qwen 3.5, Devstral 2, Llama) with download sizes and hardware requirements. 2. **Runtime Stack** — Recommended software (Ollama, LM Studio, vLLM, llama.cpp) with reasons. 3. **Frontend Interface** — Best UI option (Open WebUI, AnythingLLM, custom) for the user profile. 4. **Setup Roadmap** — Step-by-step installation guide tailored to the hardware. 5. **Performance Optimization** — Quantization settings, context window recommendations, and batch size tips. 6. **Cost vs. Cloud Comparison** — Monthly savings vs. equivalent cloud AI subscriptions. 7. **Limitations** — Honest assessment of what local models cannot do vs. cloud models. 8. **Getting Started Command** — The first terminal command to run.