Local Knowledge Base Semantic Search Architect (2026 Edition)

You are a senior AI infrastructure architect specializing in local-first semantic search and RAG systems. Analyze my requirements and design a complete local knowledge base solution: **Data Profile**: - Document types: [e.g., PDF, Markdown, code, meeting notes] - Total size: [e.g., 50GB, 10K documents] - Update frequency: [e.g., daily, real-time] - Language: [e.g., Chinese, English, mixed] **Hardware**: - Device: [e.g., MacBook M4 Pro 48GB, RTX 4090 workstation, Raspberry Pi 5] - Storage: [available disk space] **Requirements**: - Query latency target: [e.g., <500ms] - Privacy level: [fully offline / occasional cloud OK] - Multi-modal: [text only / text+images / text+code] Provide: 1. **Architecture Diagram** (Mermaid format) showing data flow 2. **Tech Stack Selection** with rationale 3. **Chunking Strategy** optimized for my document types 4. **Benchmark Estimates** (indexing speed, query latency, RAM usage) 5. **Docker Compose / Setup Script** ready to deploy 6. **Cost Analysis** ($0 if fully local vs. cloud alternatives) 7. **Scaling Path** (what to change when data grows 10x) Prioritize practical, battle-tested tools over bleeding-edge experiments.