Your meetings are dying in recordings.
Let them grow.

garden-md turns every meeting transcript into a living, browsable wiki. Entities extracted, cross-referenced, and stored as markdown files you own. No cloud lock-in. No subscription tax.

Try the playground →
$ garden-md sync --source granola
syncing 47 transcripts...
$ garden-md process
extracting entities...
● 23 people · ● 14 companies · ● 31 products
writing wiki to ./garden/
✓ wiki ready. open garden/index.html

Acme Corp

Enterprise client. Main contact: Sarah Chen (VP Eng). Discussed integration timeline in 3 meetings. Decision: Q2 pilot with 50-seat deployment.

company 3 meetings active deal
How it works

Three commands. Zero lock-in. Total recall.

01

Sync transcripts

Pull from Grain, Granola, Fireflies, or any custom source. Transcripts land in a staging area, untouched.

→
02

Extract entities

AI identifies people, companies, products, decisions. Cross-references across every meeting you've ever had.

→
03

Browse your wiki

A Wikipedia-style knowledge base, rendered as HTML. Every entity linked. Every conversation searchable. All in markdown files you control.

🔗

Wired into your coding agents

garden-md auto-configures Claude Code, Cursor, and Windsurf to access your wiki as context. Ask your AI coding assistant "what did the client say about the API?" and it knows. Your meetings become part of your development workflow.

🔒

Local-first, markdown-native

Your data never leaves your machine. Plain markdown files, version-controlled with git if you want. No vendor owns your institutional memory.

⚡

Bring your own AI

Works with Claude Sonnet, GPT-4o, Gemini Pro, or local Ollama. You pick the model, you control the cost. ~$0.01 per transcript.

Every meeting your company has ever had, organized and searchable. Finally.

Open source. MIT license. The institutional memory your team deserves, built from the conversations you're already having.

Try it now — paste a transcript →