Start with a thought. Write a question, a half-formed idea, or something you don’t want to forget. It doesn’t need a title.
01 / Shan is a Shared Human–Agent Notebook
AI Native.
Your agent runs in a purpose-built sandbox. Work with your notes the way you work with a coding agent.
02 / Shan is private
Commit your note.
One note, one file. Every change, a commit. Saved locally and synced to your own GitHub repository. Your thoughts stay yours.
03 / Shan only uses your Codex token plan
Codex Token Plan.
Shan puts your unused tokens to work behind the scenes, connecting your notes and exploring topics that spark your curiosity—all without you having to ask.
Forget Markdown. It won’t necessarily help you.
Give your curiosity a task. “Explore my notes. What’s one idea worth thinking about more deeply?”
#ASK
Follow a question over time. Add #WATCH to follow a thread. Your agent watches for news on the topic and starts a new thread when something relevant appears.
#WATCH
Capture first. Check next. Save what catches your interest. Every new note will automatically go through #CHECK: your agent checks its claims, finds sources, and flags what’s uncertain.
#CHECK
Let connections find you. When you create a note, your agent will find related threads and link them in both directions. A single thought becomes part of a bigger picture.
Search where you write. Use the bottom field to search your notes—or turn a new thought into a note.
Connect your thoughts. Type @ while writing to reference another note. Bring an earlier idea into your current thought.
Pick up an unfinished idea. “Find a topic I kept returning to but never explored deeply. Suggest a next step.”
#ASK
Keep a history of your thinking. Each note is a file. Each change is a commit. Your GitHub repository keeps the history.
Let your notes surprise you. “Find an unexpected connection between two of my notes. Explain why it might matter.”
#ASK