Search
Mindweave provides two search modes: traditional keyword search for exact matches and AI-powered semantic search that understands meaning and context.
Keyword Search
Full-text keyword search looks for exact and partial word matches across your content. It searches through:
- Content titles
- Body text
- Tags (both manual and auto-generated)
- URLs (for link content)
Keyword search is fast and useful when you know the exact terms you're looking for. Results are ranked by relevance.
Semantic Search
Semantic search uses AI to understand the meaning behind your query, returning results that are conceptually similar even when they don't share the same words.
How it works:
- Your search query is converted into a 768-dimensional vector embedding using Google Gemini.
- This embedding is compared against the embeddings of all your saved content using pgvector.
- Results are ranked by cosine similarity — the closer the vectors, the more relevant the result.
For example, searching for "machine learning algorithms" might return a note titled "Neural network training techniques" because the concepts are semantically related.
When to Use Which
| Use Case | Search Mode |
|---|---|
| Looking for a specific title or tag | Keyword |
| Finding content about a topic or concept | Semantic |
| Searching for a URL or domain | Keyword |
| Exploring related ideas | Semantic |
| Filtering by exact tag name | Keyword or Library filter |
Search Tips
- Be descriptive for semantic search — Longer, more descriptive queries give better semantic results. Instead of "react", try "react component state management patterns".
- Use keyword search for precision — When you know the exact term, keyword search is faster and more precise.
- Combine with filters — Use the library filters alongside search to narrow results by type, tags, or date.