Import and Portability

Import is useful when you already have durable project notes, meeting notes, research notes, or a Markdown vault that should become searchable AI work context.

Do not import everything just because it exists. Origin works best when imported material is useful to future agents, not when the memory layer becomes a bulk file mirror.

The vault import path reads Markdown files, chunks them, embeds them, and stores them as imported memories. The source vault is not moved or deleted.

Use a small representative vault first, then recall against it before importing a larger archive. That gives you a chance to see whether the imported context is useful or noisy.

Markdown links and wiki-style relationships are useful signal, but they are not the same as the source-backed page model.

Imported links can help graph context. Source-backed pages are stronger when Origin can trace the synthesized claim to the source memories that produced it.

If another app is your knowledge-base source of truth, keep using it for polished notes. Use Origin for the AI work loop: captures, handoffs, source-backed pages, recall, and cross-client context.

Running both is possible through MCP, but you still need a human rule for which system owns which kind of context.

Origin projects readable artifacts under ~/.origin, including pages and session handoffs. You can open those files in an editor or point a Markdown reader at them.

Plain-text access is not the same as full Origin behavior. Capture, recall, dedupe, distill, review, forget, and index updates still need the daemon and CLI/MCP paths.

Imported vaults and exported-looking Markdown can contain sensitive project history. Do not publish ~/.origin, sync it to a cloud drive, or attach it to an issue without redaction.

For backups, treat ~/.origin and the daemon data directory as private application data. If you need to recover or migrate machines, verify with origin doctor afterward so the daemon, index, and MCP config agree.