Embeddings & Semantic Search

Run a local embedding model so search ranks by meaning, not just keywords.

What This Powers

When embeddings are loaded, memrynote can rank notes by semantic similarity to your query — not just keyword overlap. This affects:

• The search palette (semantic boost on top of keyword match)
• "Related notes" suggestions in some surfaces

A query like "setting up authentication" can surface a note titled "OAuth flow" even when the words don't overlap.

Enabling

1. Open Settings → AI
2. Toggle Enable
3. Under Embedding Model, click Download to pull the model
4. Wait for the status to say Loaded
5. Click Rebuild Index to embed every existing note (one-time per model)

The first index build can take a few minutes for large vaults — progress is shown.

Model Management

The status line shows:

• Loaded — ready
• Loading — initialization in progress
• Not downloaded — needs download
• Error — see logs; usually disk space, a hash mismatch, or a failed download

You can Unload the model from settings to free memory; reload as needed.

When the Download Fails

The model is fetched once (~23MB). If that download fails — you are offline, behind a proxy, or the CDN is blocked — memrynote does not hammer the network. It waits before trying again, backing off each time (about one minute, then two, four, and eight), and after several consecutive failures it stops retrying for the rest of the session. Semantic search falls back to keyword-only meanwhile; nothing else is affected, and no notes are lost.

If the connection comes back on its own, a later retry picks it up and indexing resumes with no action from you. To retry immediately instead of waiting out the backoff, do any of these — each one clears the wait:

• Toggle Enable off and on in Settings → AI
• Click Download / Load model
• Click Rebuild Index

Restarting memrynote also clears it.

Opening a vault never blocks on embeddings. When a vault has notes that still need embedding — for example the first open after importing a vault — memrynote embeds them in the background after the vault is already open, so a large vault (or a slow or failed model download) can never hold up opening.

Closing a vault, switching vaults, and quitting never block on embeddings either: a background embedding pass stops at the next note rather than finishing its whole queue. Notes it did not reach keep their place in line and are embedded by the next background pass.

Beyond that, the model is loaded lazily: semantic surfaces such as search, inbox linked-note suggestions, related notes, and reindexing start the local model on first use. The model runs in a separate utility process and shuts down after an idle period, so regular note reading does not keep the embedding runtime resident forever.

Model Size

Models trade off accuracy vs disk and memory. The default is tuned for desktop hardware. The settings page shows dimensions and the current count of embedded notes.

Reindexing

Rebuild the index after:

• Switching models
• Restoring a vault from backup
• A migration that touched note storage

Reindexing is incremental — memrynote skips notes whose content hash hasn't changed.

A reindex — and a settings change that reclassifies notes, such as moving the journal or default note folder — embeds the notes it touched in the background as soon as the pass finishes. You do not need to reopen the vault for semantic search to see them.

Privacy

Embeddings are computed on-device. The vectors are stored in the local index database (<vault>/index.db). They are never sent to a server.

Even if you sync across devices, embeddings are recomputed locally — the embedding payload itself is not part of the sync stream.

Performance

Once the index is built, semantic search adds <50ms to a typical query. Embedding is the expensive step (one-time per note); ranking is cheap.

If you have an enormous vault and notice slowdowns, the index can be rebuilt fresh in settings.

Disabling Embeddings

Toggle Enable off. The model unloads. The vector index stays on disk (you can delete the file manually if you want it gone).

Search falls back to keyword-only — fast, but less forgiving of varied phrasing.

See Also

• Search & Command Palette
• Provider Setup — provider config for the inline AI menu (separate from embeddings)