The Harvest method
Sales notes, research transcripts, decision memos, support tickets — your team produces more written evidence in a quarter than any human can read in a year. Harvest is the methodology for compounding it. Winnow is the open-source tool that runs it.
"Harvest" can pattern-match to data-harvesting — the surveillance-flavoured kind, where someone gathers up signals about your customers without their knowing. That isn't what this is. The object of the verb here is your team's own writing — sales notes, research transcripts, decision memos — first-party evidence you authored and already own. It's a farmer harvesting their own crops, not someone helping themselves to the field next door.
The full version, including what enters the system, what doesn't, and what flows out, lives at where the data comes from, where it goes.
How it works
Drop the things you already write into a folder. The wiki maintains itself. Every output cites the evidence it came from.
Sales calls, support tickets, user research, competitor scans, system exports, decision memos. Markdown, PDF or DOCX. Drag-and-drop, webhook, or a cloud-synced folder — Winnow doesn't care which path it takes. Raw evidence is immutable: once it's in, it stays in, untouched, as a source of truth.
An LLM reads each new piece of evidence, decides which wiki pages should change, and proposes the edits. Reinforcement and contradiction with prior beliefs are flagged explicitly — no silent overwrites. You review the diff before anything writes.
Discover opportunities from tension across the wiki. Generate recommendations only when a validated opportunity exists. Hand them straight into Claude Code, Cursor or Lovable — every brief carries the evidence it came from and the foundations as non-negotiable guardrails.
Winnow is open-source and free to run on your own infrastructure. The hosted version at usewin.now is opening in waves — join the waitlist to get early access.