The three-phase cycle
Thoughtbed doesn't ask you to learn a system. It asks you to keep doing what you're already doing — capturing thoughts — and lets the engine handle the rest.
Plant
Drop raw captures, voice memos, or browser clips into the bed. No folders, no tags, no friction. Just capture. Every note becomes a seed — timestamped, contextualized, and waiting.
Grow
The engine connects your notes while you sleep, finding structural patterns across months of thinking. It doesn't search for keywords — it looks for convergence. When three unrelated notes from three different contexts start pointing at the same idea, the engine notices.
Harvest
Get notified when a cluster of thoughts reaches critical mass. When it's ripe, it's ready to become something real — a newsletter, a product, a conversation, a decision. The system delivers it with proof of why it's ready.
The scoring formula
Every idea in your Thoughtbed is scored on three dimensions. No subjective rankings. No manual tagging. Just math applied to your own thinking patterns.
Breadth — How many parts of your life is this idea touching? Work, personal, creative, reading, conversations. The more domains, the more load-bearing the idea.
Depth — How many independent sources pointed to it? A podcast, a book, a shower thought, a client call. Each source is a confirmation the idea isn't noise.
Momentum — Is it still growing? An idea that attracted three sources last week and another one today is alive. An idea that peaked six months ago is dormant.
The delivery
When an idea crosses the threshold, you receive a Thoughtbed Dispatch — a delivery card that tells you exactly what's ripe and why.
Each dispatch includes the idea's scores, the sources that contributed to it, the life domains it spans, and how long it's been growing. No guesswork. No "you might also like." Just your own thinking, matured and ready.
Search is reactive — you have to know what you're looking for. Maturation is proactive — the system tells you what's ready before you even think to ask.
What it runs on
Thoughtbed runs locally on your machine. Local LLMs. Local vector databases. Your private thoughts stay private — nothing leaves your device.
It works with any app that stores data as Markdown files — Obsidian, Logseq, Roam via export. Connect your vault, install the engine, and keep writing where you already write.
Your first Harvest report arrives in 7 days.