Build Private AI Knowledge / “Second Brain” System — Obsidian + n8n + RAG + Voice Capture

I am looking for an experienced n8n/RAG developer to build a private AI-powered knowledge and intellectual-property system. The purpose is to help me capture, organize, connect, search, and develop ideas for an original professional-development framework, book, course, executive coaching methodology, workshops, and thought leadership. This is not a generic chatbot, marketing automation, CRM, or social-media project. I already have a detailed architecture specification and knowledge model. I need someone who can implement it. The system should include: * Obsidian/Markdown as the permanent, canonical knowledge repository * Structured records for Sources, Concepts, Original Insights, Stories, Research/Evidence, Frameworks, Outputs, and Inbox * YAML metadata and stable record IDs * n8n-based automation * Voice-note capture and automatic transcription * Text and document/PDF ingestion * Automatic classification and tagging * Suggested relationships between related ideas * Clear provenance tracking distinguishing external source material, my interpretation, my original IP, empirical research, and AI-generated synthesis * Vector database and semantic search/RAG * Ability to conversationally query the entire knowledge base * Links/references back to underlying Obsidian records * Vector index that can be completely rebuilt from the Markdown knowledge base * Cost-conscious, configurable LLM/model selection * User-owned VPS, APIs, databases, credentials, and infrastructure * Backup/export capability * Complete workflow, prompt, configuration, and handoff documentation One initial capture channel such as Telegram or WhatsApp is sufficient for Version 1. The architecture should make additional channels easy to add later. Example workflow: I send a voice note containing an idea or professional story. The system transcribes it, identifies whether it is a Story, Original Insight, Source Note, etc., assigns a stable ID, creates a structured Markdown record, suggests relevant concepts/tags/relationships, indexes it for semantic retrieval, and confirms what was captured. Later I should be able to ask: “Show me everything in my knowledge base related to why successful executives sometimes become trapped by their existing success. Include my original insights, stories, relevant source thinkers, research evidence, contradictions, and any framework principles I’ve developed.” The system should retrieve and synthesize the relevant material while preserving provenance. Ideal experience: * Advanced n8n * RAG / semantic search * Vector databases such as Qdrant, Weaviate, Pinecone or pgvector * OpenAI/Anthropic APIs * Obsidian/Markdown/YAML * Python and/or JavaScript * Docker/Linux/VPS deployment * Speech-to-text APIs * Telegram or WhatsApp API integration * Knowledge architecture/data modeling Important: Obsidian/Markdown must remain the source of truth. The vector database should be a disposable/rebuildable semantic index rather than the only location where knowledge is stored. I have a detailed technical specification that will be provided to qualified candidates. When applying, please provide: 1. Examples of similar RAG/knowledge-management systems you have personally built. 2. Your recommended architecture for this project. 3. Which vector database you recommend and why. 4. How you would ensure the vector index can be rebuilt from the Markdown repository. 5. How you would preserve provenance between external sources, my original ideas, and AI-generated synthesis. 6. Your fixed-price estimate and delivery timeline. 7. Post-delivery support included. 8. Estimated ongoing monthly infrastructure/API costs. Please begin your response with “IP Brain” so I know you have read the project requirements.