# Case Study: Obsidian Startup OS > **A personal knowledge operating system for AI-era creative practitioners** --- ## Summary **EN:** The Obsidian Startup OS is HelenQ's personal knowledge management system — designed not as a simple note-taking setup, but as a full cognitive infrastructure for an AI-augmented creative practice. It treats the individual creative practitioner as a "startup of one," requiring the same knowledge architecture that a growing organization would need: interconnected project knowledge, retrievable institutional memory, a living SOP library, and AI-ready structured data. **ZH 摘要：** Obsidian Startup OS 是 HelenQ 的个人知识管理系统——不是简单的笔记设置，而是为 AI 时代创意实践设计的完整认知基础设施。它将独立创意从业者视为"一人创业公司"，需要与成长型组织相同的知识架构：相互关联的项目知识、可检索的机构记忆、动态 SOP 库以及 AI 就绪的结构化数据。 --- ## Project Metadata | Field | Detail | |---|---| | **Project Type** | Personal knowledge OS / Workflow infrastructure | | **Time** | 2024 – Present (continuously maintained) | | **My Role** | Sole designer and user | | **Status** | Active / Core to daily practice | | **Primary Tool** | Obsidian (with selected plugins) | --- ## Problem Creative practitioners in the AI era face a specific knowledge management challenge that generic productivity tools don't solve: 1. **Knowledge fragmentation:** Insights, prompts, workflows, project learnings, and client context live across chat threads, documents, and memory — never integrated 2. **AI context poverty:** When using AI tools, lack of structured personal context means AI outputs are generic — not informed by your actual expertise and history 3. **Workflow amnesia:** Workflows discovered and refined through effort are not captured — so each project re-invents from scratch 4. **Scaling without a team:** Independent practitioners cannot hire knowledge managers or librarians — they need a system that builds institutional memory automatically through normal work patterns 5. **The "second brain" trap:** Most PKM (Personal Knowledge Management) systems become graveyards of saved content with no active retrieval or use The failure of the $2M AI knowledge base project (documented in Project 03) was partly attributable to these same issues at organizational scale — knowledge architecture designed without understanding actual knowledge retrieval needs. --- ## Solution An Obsidian vault structured around five interconnected layers: ### Layer 1 — Project OS Each active project gets a structured "Project Hub" note: - Brief summary and current status - Key decisions log (what was decided, why, by whom) - AI conversation links (direct links to key AI chat threads) - Deliverable tracker - Learnings capture (auto-prompted at project close) ### Layer 2 — SOP Library Living documentation of every workflow HelenQ has systematized: - Tool-specific SOPs (how to run a ComfyUI session, how to structure a Gemini prompt for a pitch) - Template library (reusable prompt templates, document structures) - Version history (SOPs update as tools and methods evolve) - Tagged by domain, tool, and output type for fast retrieval ### Layer 3 — Prompt Vault Structured storage for prompts that work: - Categorized by: tool, content type, style, purpose - Each prompt entry includes: the prompt, example outputs, notes on what it does/doesn't do, last tested date - Enables retrieval of "what's my best current approach to X" in seconds ### Layer 4 — Knowledge Graph Connected notes for cross-domain synthesis: - Books, articles, and ideas connected to projects and SOPs that use them - People and companies connected to projects and relationship context - Tools connected to SOPs, prompts, and projects that use them - Enables: "show me everything connected to AI knowledge systems" as a usable cluster ### Layer 5 — AI-Ready Structured Data Notes formatted for AI consumption: - Personal context documents (my positioning, my working style, my active projects) — fed to AI sessions for personalized context - Structured project briefs that can be directly pasted into AI conversations - "Who am I for this conversation" templates for different AI use contexts (strategic, creative, technical) --- ## Tools & Plugins - **Core:** Obsidian (local, private, Markdown-based) - **Plugins:** Dataview (database queries across notes), Templater (auto-structured new notes), Canvas (visual mapping), Daily Notes (structured daily capture) - **Backup:** iCloud sync + periodic Git backup - **AI integration:** Notes formatted for direct copy-paste into AI conversations; periodic use of AI to process new inputs into structured vault entries --- ## Deliverables 1. **Vault architecture** — folder structure, naming conventions, and linking standards 2. **Note templates** — Templater templates for Project Hub, SOP, Prompt, Research Note, Daily Note 3. **Dataview queries** — pre-built queries for: active projects, recently updated SOPs, prompts by tool, weekly review 4. **AI context documents** — ready-to-use personal context files for AI session initialization 5. **Maintenance protocol** — weekly review ritual and capture habits that keep the system alive --- ## Result - Active knowledge base with interconnected project, SOP, and prompt content - AI sessions are noticeably more personalized and contextually accurate when initialized with vault context documents - Reduction in "I know I figured this out before but can't find it" moments - System has informed the design of the WOAI SOP framework and the Orbit editorial system (both documented separately) - Serves as the "brain" behind the independent practice — tracking learnings, decisions, and knowledge across all projects *Note: This is a personal system with no external user metrics. Results are practitioner-reported.* --- ## Reusable Value The Obsidian Startup OS represents a design pattern for **AI-era personal knowledge infrastructure**: 1. **Treat yourself as an organization:** Your knowledge needs the same architecture that a company's knowledge base needs — retrieval-optimized, not just storage-optimized 2. **AI-readiness is a design goal:** Structure your knowledge so AI can consume it directly — not just so humans can read it 3. **SOPs and prompts are first-class knowledge objects:** Not just "notes" — they deserve dedicated structure, versioning, and retrieval 4. **The system must earn its maintenance:** If the system is harder to maintain than the value it provides, it will die. Design for minimal friction capture with maximum retrieval value --- ## Related Context This system directly addressed lessons from the $2M AI Knowledge Base failure (Project 03 Autopsy): - The failed project tried to build organizational knowledge infrastructure without understanding how knowledge is actually retrieved and used - The Obsidian OS is designed around actual retrieval patterns — what do I need to find quickly? Under what circumstances? — not around comprehensive storage --- ## Keywords personal knowledge management, PKM, Obsidian, second brain, knowledge architecture, SOP library, prompt management, AI context, creative workflow, independent practitioner, knowledge OS, Dataview, Templater, AI-augmented practice --- *HelenQ Case Study | June 2026*