Setup
Where Wenlan Stores Claude Code Memory
Find the local files Wenlan writes when Claude Code captures memories, handoffs, and distilled pages.
Article packet
Workflows
Claude Code users checking where memory lives on disk
5 min read
01
Wenlan writes readable artifacts under ~/.wenlan instead of hiding memory inside a chat profile.
02
Distilled pages live under ~/.wenlan/pages and session handoffs live under ~/.wenlan/sessions.
03
Readable pages and session handoffs are versioned in ~/.wenlan/.git.
01
Quick answer
Wenlan stores its local runtime and readable artifacts under ~/.wenlan. That is separate from Claude Code's own CLAUDE.md and auto-memory systems: Claude Code runs the plugin and calls MCP tools, while Wenlan's daemon owns the local memory store.
Wenlan's answer is deliberately boring: the memory layer is local. Memories live in the daemon database, while pages, handoffs, project status, and local git history are exposed as readable artifacts.
02
When this problem appears
This question usually comes up after a user installs the Claude Code plugin and wants to know whether memories are trapped inside Claude Code, stored in the cloud, or written somewhere they can inspect.
03
Inspect the local artifacts
After setup, use the filesystem and CLI to confirm what Wenlan wrote.
- Check ~/.wenlan/pages for distilled Markdown pages.
- Check ~/.wenlan/sessions for handoffs and project status.
- Check ~/.wenlan/db on macOS or Linux for the convenience symlink to the libSQL store.
- Run ~/.wenlan/bin/wenlan status to verify the daemon.
- Inspect ~/.wenlan/.git when you need history for readable artifact writes.
Local artifact checks
ls ~/.wenlan
ls ~/.wenlan/pages
ls ~/.wenlan/sessions
~/.wenlan/bin/wenlan status
git -C ~/.wenlan log --oneline -504
What to check next
Do not edit the database directly. MCP memories live in the daemon DB; distilled pages and session handoffs are the readable projection. Use Wenlan commands and tools for memory writes, review, distill, and delete.
Try the local memory loop
Install Wenlan, connect your AI client, and verify that capture, recall, and handoff work on your machine.
FAQ