A guide to shared memory

Shared memory. A clear path.

Your guide to Personal Agentic OS, from choosing tools to your first sync.

What are you looking for?

Search runs on this device.

On this page

Learning path

Build local shared memory when you work across projects and AI tools.

Prepare Node.js, the destination path, and your Obsidian decision before installation.

Run the local installer, review its proposed plan, and approve changes only afterward.

Choose Obsidian for shared memory or knowingly accept Dashboard-only limitations.

Select Codex, Claude Code, or Gemini CLI and approve the proposed local paths and permissions.

Check connections, the proposed project map, and sync results before relying on the Dashboard.

Use one reference for project decisions and current state without copying entire conversation histories.

Separate tool instructions from operational memory and avoid duplicated or conflicting rules.

A controlled handoff of decisions, sources, and open questions—not an entire conversation export.

A Markdown file for goals, state, decisions, and next actions, with a synthetic example and usage rules.

Understand the boundaries between your Vault, model provider, and website analytics; local files do not make every workflow offline.

Keep project identity, provenance, and context packs separate; review proposed merges before approval.

An actual 0.10.0 synthetic-data result: adapter creation, a context pack, and verification without saved-state changes.