This is how to build an AI memory system

Obsidian AI Memory System Graph view - my-vault II NOTES LOCAL FOLDER how AI system OWNERSHIP 60 SECONDS You Actually Own It No database. No account. Just files on your disk. FREE & OPEN SOURCE PLAIN MARKDOWN Nobody can take it away. PORTABLE NO LOCK-IN Switch In Minutes Models change. Prices change. Your files don't. my-vault /~/Vault/AI-Memory/ README.md what this project is STATUS.md where it stands today PROGRESS.md what got done, when DECISIONS.md why we chose it START YOUR VAULT FREE & OPEN SOURCE Claude $1/B tokens ChatGPT $0.50/B tokens Gemini $0.25/B tokens CONNECTED READY my-vault /~/Vault/AI-Memory/ README.md PROGRESS.md DECISIONS.md STATUS.md START YOUR VAULT FREE & OPEN SOURCE models changes over minutes Sharpen your thinking. The free and flexible app for your private thoughts. Get Obsidian for Windows open notes telepathy obsidian free Cloud Back Create local vault Vault name Pick a name for your awesome vault. Location Pick a place to put your new vault. Browse Create Nathan's AI memory Browse Your new vault will be placed in: D:\client\12 June\Nathan\Obsidian AI Memory System\Nathan AI memory Create want folder project to any that write files to starter builds Content startup-prompt README AI Content Pipeline What this project is My content production system for short-form video and posts about AI tools, AI updates, and automation workflows. Covers everything from spotting what's worth covering through to a filmed, edited, published piece. Who it's for Creators, founders, and people who want to keep up with AI without reading release notes all day. I publish on Instagram, LinkedIn, and YouTube. Why it matters The bottleneck isn't filming. It's deciding what's worth covering and turning it into a script fast. Most AI news is noise. The pipeline exists to filter signal and get it to camera quickly, while it's still current. Format 9:16 vertical, 30-60 seconds Content: README Status STATUS Status - AI Content Pipeline Last updated: 2026-07-25 Where things stand Mid-production on the Obsidian AI memory system video. Script is locked at 8 lines (~40 seconds). Recording guide is written with a shot-by-shot breakdown. Vault and memory files are set up and populated. Now moving into filming. Done Script locked - 8 lines - ~40 seconds Recording guide written (Windows) Four prompt files written and tested Obsidian vault created and project folder setup Memory files populated with real content Still open Record voiceover in one clean pass Shoot Batch 1-2 (browser + Obsidian setup) Shoot Batch 3 (Claude session - must run in order) Build Canva text cards Edit and cut to 40 seconds Decide on the CTA line Content: progress Progress Progress Log - AI Content Pipeline Newest entries at the top. 2026-07-25 Worked on staging the memory files with real content ahead of filming. Changed: Replaced blank templates with populated versions so the on-camera file reveals actual content. Tried: Running the session-start prompt against empty template files. Didn't work Returned: