Index - Project Requirements Document Vault
Source: /Users/nitishchauhan/Downloads/Project Requirements Document - Vault.pdf
Type: product-spec
Status: indexed
1. Recurring keywords / key ideas
- Vault (working title) / export-first utility
- Cognitive closure / close the mental loop
- Capture First, never summarize or rewrite
- Universal AI chat platform support
- Sidebar temporary workspace and multi-item queue
- Metadata, asset capture, local buffer retry
- Explicit V1 out-of-scope (RAG, embeddings, knowledge graphs, Substack generation)
- Version 1.5 optional 48-hour temporary cloud vault
- React / TypeScript / Tailwind / MV3 / JSZip / Turndown / DOMPurify
- Guiding question: capture loop vs knowledge management creep
2. Core / novel / atypical ideas
- Vault is not a second brain, KM system, or AI memory platform; it is export-first cognitive closure
- Problem framed as tabs kept open from fear of forgetting unfinished mental loops
- Philosophy: interpretation optional, capture mandatory; extension never interrupts thinking
- Design principles: Capture First, User Control, reduce mental effort, Universal Support, Modular Architecture
- Primary flow ends when mental loop is closed after queued multi-site export
- Harsh V1 exclusions (summaries, embeddings, RAG, LinkedIn/Substack generation, recommendations) to protect product identity
- V1.5 temporary cloud vault with 48-hour auto-deletion to avoid digital hoarding while adding safety
- Future modules (MCP, self-hosted archive, plugins) must stay independent; capture never depends on AI processing
- Success includes reducing need to leave tabs open from fear of loss
- Feature gate question: does this help capture and close the loop, or turn Vault into another KM system?
3. Context and flow
The PRD names the Chrome/WebExtension product Vault and scopes it tightly around frictionless conversation capture and export across many AI platforms. It states vision and anti-vision, catalogs user problems, sets design principles, defines primary flow and V1 features (universal capture, sidebar, queue, export formats, metadata, assets, local buffer, generic webpage capture), then fences out AI-heavy futures. Technical direction and success criteria reinforce that reliability and loop-closure beat knowledge-management ambition.
4. Publishing angles
- Mode: Authority article
- Why: Clear product thesis with anti-scope and a durable decision rule against KM feature creep.
- Angles:
- Cognitive closure as the real job of an AI chat export tool
- Why Vault refuses to be a second brain in V1
- The tab-hoarding problem: fear of losing conversations as UX debt
- A product rule for scope: capture loop vs knowledge management