--- mw_bundle: 1 id: 4OGRyHs9 title: "Memory.Wiki in practice: case studies" url: https://memory.wiki/b/4OGRyHs9 document_count: 4 updated: 2026-05-24T17:48:04.170Z analysis_generated_at: 2026-05-11T03:33:09.941Z analysis_stale: true source: "memory.wiki" --- # Memory.Wiki in practice: case studies > Four short stories about how Memory.Wiki actually gets used: cross-tool handoff, CLAUDE.md onboarding, sharing AI conversations like Notion links, and the personal LLM wiki Karpathy described automated. > ⚠ _Analysis may be stale — one or more member docs were edited after the last analysis run. Re-run the canvas to refresh._ ## Summary These documents demonstrate mdfy's approach to solving AI workflow inefficiencies through automated knowledge capture and sharing. They showcase four key use cases: eliminating repetitive context-setting in fresh AI sessions, maintaining decision continuity across different AI tools, automating personal knowledge base maintenance, and creating shareable permanent URLs from AI conversations. The collection positions mdfy as a bridge between ephemeral AI interactions and persistent, searchable knowledge assets. ## Themes - Context persistence across AI sessions - Multi-tool workflow integration - Automated knowledge management - Collaborative AI conversation sharing ## Cross-document insights - The real bottleneck in AI-assisted work isn't the AI's capabilities, but the human overhead of repeatedly establishing context and transferring knowledge between tools and sessions. - mdfy represents a shift from manual knowledge curation (like Karpathy's approach) to selective automation where humans approve what enters the knowledge base but machines handle organization and cross-referencing. - The documents reveal a tension between vendor-specific AI tool ecosystems and the need for cross-platform knowledge persistence - mdfy serves as vendor-neutral infrastructure. - The emphasis on permanent URLs that work for both humans and AI suggests a future where knowledge sharing protocols are designed for hybrid human-AI consumption rather than just human readers. ## Key takeaways - mdfy eliminates the repetitive overhead of re-establishing context in AI conversations through one-time setup and permanent URLs - The platform enables seamless knowledge handoff between different AI tools, maintaining decision continuity across vendor boundaries - It automates 80% of personal knowledge base maintenance while keeping humans in control of what gets captured and shared ## Open questions / gaps - No discussion of data privacy, security, or enterprise compliance considerations for storing AI conversation data - Missing details about pricing, scalability limits, or technical infrastructure requirements - Lacks comparison with other emerging AI memory/knowledge management solutions beyond brief mentions of Notion and Obsidian - No analysis of potential failure modes, such as what happens when the mdfy service is unavailable or URLs become broken ## Notable connections - **doc:case-claude-md-personal-context** ↔ **doc:case-cross-tool-handoff** — Both solve context persistence but the second extends to multi-tool scenarios while the first focuses on single-tool optimization - **doc:case-cross-tool-handoff** ↔ **doc:case-share-with-team** — The handoff document establishes individual workflow benefits that the sharing document extends to team collaboration scenarios - **doc:case-personal-llm-wiki** ↔ **doc:case-claude-md-personal-context** — The wiki document provides the theoretical framework and architecture that the onboarding document implements as a practical application - **doc:case-share-with-team** ↔ **doc:case-personal-llm-wiki** — Both emphasize permanent URLs and knowledge persistence but sharing focuses on collaboration while wiki focuses on personal knowledge management automation ## Concepts (this bundle) - **Context Persistence** - **One-Time Setup** - **Knowledge Handoff** - **Decision Continuity** - **Automation vs Curation** - **Permanent URLs** - **Dual Format Access** - **Capture Workflow** ## Concept relations - **Context Persistence** ↔ **One-Time Setup** — solved by - **Knowledge Handoff** ↔ **Decision Continuity** — enables - **Automation vs Curation** ↔ **Capture Workflow** — influences design - **Permanent URLs** ↔ **Dual Format Access** — implements ## Documents ### 1. [Cursor for code, Claude for research, finally on the same page](https://memory.wiki/case-cross-tool-handoff) Install `/memory.wiki` once takes 30 seconds via a single curl command · Per-capture overhead is under 5 seconds · Per-conversation context load is roughly 3 seconds (paste the hub URL) · Same hub URL works across Cursor, Claude Code, and any other AI tool · Verified across Claude, OpenAI, and Gemini at 100% accuracy via [MWBench](/mwbench) *sections:* The pain: I do real coding in Cursor. I do research and architecture thinking in Claude. They don't share memory. Every time I switch, I re-paste the same context: "here' | What I do now: In Cursor, after we figure out a non-trivial decision: save this to memory.wiki as "Auth provider tradeoffs". The skill captures the conversation segment as a p | What changed: The handoff is one capture command, one paste.; I never re-explain the project to a fresh chat.; The decision history is searchable as one knowledge layer, not split across vendor walls. | … ### 2. [Onboarding every new Claude Code session in one line](https://memory.wiki/case-claude-md-personal-context) A developer can eliminate repetitive project context setup across Claude Code sessions by adding a single Memory.Wiki hub URL reference to their ~/.claude/CLAUDE.md file, allowing Claude to automatically fetch relevant project details on demand instead of re-explaining them each time. This same approach works across other AI coding tools like Cursor and ChatGPT. *sections:* The pain: Every time I open a new Claude Code session in a different repo, or even after a long break, I'd spend 5 minutes re-pasting the same project background. "We use | What I did: Visited memory.wiki/install, signed in, copied the snippet, appended it to ~/.claude/CLAUDE.md: | Personal context (Memory.Wiki hub): Hub URL: https://memory.wiki/hub/ | What changed: Now every Claude Code session, anywhere on my machine, automatically has my hub URL. When I ask a question that needs project context, Claude fetches the hub an | Time: One-time setup: 1 minute.; Per new chat: 0 minutes. The hub URL is already in CLAUDE.md. | … ### 3. [Sharing AI conversations the way you'd share a Notion link](https://memory.wiki/case-share-with-team) A tool called Memory.Wiki lets users save Claude conversations as permanent shareable URLs that work for both human readers (with formatting) and AI systems (as raw markdown), solving the problem of sharing AI chat context with teammates without reformatting or losing functionality. *sections:* The pain: I had a really useful Claude session about our deploy pipeline. Three colleagues needed to read it. The options were:; Screenshot the chat: terrible to skim, breaks links.; Copy and paste the markdown: loses code highlighting, no permanent URL.; Send the Claude share link: only works for ChatGPT or Claude users with the right account, breaks if I edit settings.; Re-explain in Slack: 30 minutes I don't have. | What I do now: In Claude Code: save this to memory.wiki as "Deploy pipeline review". Get back https://memory.wiki/abc123. Paste in Slack. | … ### 4. [Karpathy's hand-curated LLM wiki, without the hand-curation](https://memory.wiki/case-personal-llm-wiki) Memory.Wiki automates 80% of the maintenance burden in Karpathy's hand-curated personal LLM wiki system by automatically capturing content, synthesizing wiki pages, building semantic graphs, and surfacing gaps, while keeping humans in the loop only for accepting proposed wiki updates. *sections:* The pain: [Andrej Karpathy described the problem](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f): keep a personal LLM wiki of the answers worth rememb | What Memory.Wiki does: Same shape as Karpathy's. Different effort: | What this case actually replaces: If you're already running a personal Notion, Obsidian, or DEVONthink for AI outputs and feel the hand-curation tax, Memory.Wiki is the lower-effort version of t | … _Digest view — follow any link above to fetch that doc's full markdown. Add `?full=1` to this URL for the concatenated payload._