-
Notifications
You must be signed in to change notification settings - Fork 0
Settings
Access via Settings → Community Plugins → Vaultend.
| Setting | Options | Default |
|---|---|---|
| Display Language | Auto (follow Obsidian) / English / Korean | Auto |
Views update immediately when changed. Command palette names update after restarting Obsidian.
| Setting | Description | Default |
|---|---|---|
| AI Provider | Select the AI service: OpenAI, Google Gemini, Ollama (Local), or Custom (OpenAI-compatible) | OpenAI |
| API Key | Your provider's API key (not needed for Ollama) | — |
| Model | Select from the dropdown or choose "Custom" to enter any model ID | gpt-4o |
| Embedding Model | Model for semantic similarity (tag matching, link suggestions, duplicate detection) | Auto-detected per provider |
Changes apply immediately — no restart needed.
Ollama (Local):
| Setting | Description | Default |
|---|---|---|
| Ollama Base URL | URL of the Ollama server | http://localhost:11434 |
Ollama runs models locally — no API key needed. Install models with
ollama pull <model>. For embedding features, also install an embedding model (e.g.,nomic-embed-text).
Custom (OpenAI-compatible):
| Setting | Description | Default |
|---|---|---|
| Custom Base URL | Base URL for the API endpoint | — |
| Custom API Key | API key for the endpoint (optional) | — |
| Custom Model | Model name to use | — |
| Provider | Chat/Completion | Embeddings |
|---|---|---|
| OpenAI | Yes | Yes |
| Google Gemini | Yes | Yes |
| Ollama (Local) | Yes | Yes (requires embedding model) |
| Custom | Yes | Depends on endpoint |
Embedding-based features include: tag similarity ranking, cross-language tag resolution, and link suggestions. These features are degraded without embedding support.
| Provider | Default Embedding Model |
|---|---|
| OpenAI | text-embedding-3-small |
| Google Gemini | gemini-embedding-001 |
| Ollama | nomic-embed-text |
Organize modals and folder results show token usage and estimated cost after each AI call.
| Item | Accuracy | Source |
|---|---|---|
| Token count | Exact | Returned by the AI provider's API response |
| Estimated cost | Approximate | Calculated from a built-in pricing table |
Why cost is approximate:
- The pricing table is updated with each plugin release, but may lag behind provider price changes
- Cached-token discounts, batch discounts, and free-tier credits are not reflected
- Models not in the table show "Cost: N/A" instead of an incorrect estimate
- Ollama (local) always shows $0.00 (no API cost)
For exact billing, check your provider's usage dashboard:
- OpenAI: platform.openai.com/usage
- Google AI Studio: aistudio.google.com
| Setting | Description | Default |
|---|---|---|
| Auto-apply results | When enabled, AI results are applied immediately without manual review | Off |
| Max Response Tokens | Maximum tokens for AI response (1024–16384). Higher values cost more. | 8192 |
| Setting | Description | Default |
|---|---|---|
| Similarity Threshold | Minimum cosine similarity for embedding-based link suggestions (0.30–0.80). Lower values suggest more links but may include less relevant ones. | 0.40 |
| Setting | Description | Default |
|---|---|---|
| Auto Maintenance | Run vault maintenance periodically | Off |
| Maintenance Interval | Minutes between automatic scans | 60 |
| Reject Suppression (days) | Days to suppress a rejected proposal before showing it again | 7 |
| Exclude Folders | Folders excluded from maintenance scans (chip UI with vault autocomplete) | — |
| Exclude Tags | Notes with these tags are skipped during scans (chip UI with vault autocomplete) | — |
| Archive Folder | Destination folder for archived notes | Archive |
Maintenance scans only Markdown (
.md) notes. Non-text files (Excalidraw, Canvas) are automatically excluded.
| Setting | Description | Default |
|---|---|---|
| Embedding weight | Weight multiplier for embedding results in hybrid search. Higher values favor semantic similarity over keyword match. | 4.0 |
| RRF K parameter | Reciprocal Rank Fusion smoothing parameter. Lower values make rank differences more dramatic. | 20 |
See Privacy Rules for detailed configuration guide.
| Setting | Description |
|---|---|
| Privacy Rules | List of rules controlling what data is sent to AI. Each rule has a name, type, pattern, and enable/disable toggle. |
Vaultend stores internal data in .vaultend/ inside your vault:
| File | Purpose |
|---|---|
search-index.json |
BM25 keyword search index |
embeddings.json |
Vector store for note embeddings |
tag-embeddings.json |
Tag embedding cache |
note-embeddings.json |
Note embedding cache |
tag-groups.json |
Tag group cache (Organize Tags) |
tfidf-corpus.json |
TF-IDF corpus stats |
dirty-set.json |
Changed files tracker (smart scheduling) |
history/ |
Activity log entries |
This folder is automatically created and managed by the plugin. You generally don't need to interact with it directly.
Note: If you move or sync your vault, the
.vaultend/folder should be included. Embedding caches may need to regenerate if the AI provider changes.
Vaultend Wiki
Features
Reference