Oyma delivers sub-10 millisecond full-text search across vaults containing over 10,000 notes. By utilizing an embedded SQLite FTS5 search index, queries execute instantaneously as you type without UI stutter or latency.
High-Performance Local Indexing
As a personal knowledge base grows from hundreds of notes to tens of thousands of documents accumulated over years, naive text scanning approaches (like linear file reading) grind to a halt. Opening a search palette and waiting two seconds for results destroys intellectual flow.
Oyma is engineered specifically to maintain sub-10 millisecond search response times, even in massive enterprise-scale vaults.
The SQLite FTS5 Full-Text Engine
To achieve instant search across massive archives, Oyma maintains a local full-text search index powered by SQLite’s battle-tested FTS5 (Full-Text Search 5) extension:
[Keystroke in Quick Open: "distributed consensus"]
|
v (< 2ms)
[Local SQLite FTS5 Porter Stemmer Index]
|
v (< 4ms)
[BM25 Ranking & Token Scoring]
|
v (< 1ms)
[Render Top 20 Results in UI]- Inverted Indexing: Words are indexed into high-speed B-trees with BM25 algorithmic relevance ranking.
- Prefix Matching: Search queries match prefixes instantly as you type each letter.
- Token Stemming: Searches for “connect” automatically match “connection”, “connected”, and “connecting”.
- Zero Cloud Network Latency: Because the database resides on local NVMe storage, search speed is bounded only by local memory and CPU performance.
Benchmark Performance Metrics
Testing conducted on an Apple Silicon M2 MacBook Air with a vault containing 12,500 Markdown files (totaling 8.4 million words) yielded the following average latencies:
| Operation | 1,000 Notes | 10,000 Notes | 25,000 Notes |
|---|---|---|---|
Filename Quick Open (⌘O) | < 1 ms | 2.1 ms | 4.8 ms |
| Exact Phrase Full-Text Search | 1.8 ms | 6.4 ms | 12.2 ms |
| Fuzzy Title Search | < 1 ms | 3.5 ms | 7.9 ms |
| RAM Footprint of Index | 14 MB | 42 MB | 98 MB |
Regardless of how large your knowledge repository becomes, Oyma ensures that retrieval remains instantaneous.
High-Speed Indexing Across Massive Note Repositories
Scaling search across tens of thousands of documents:
- SQLite FTS5 Architecture: Full-text queries execute in sub-10 milliseconds against local NVMe storage.
- BM25 Relevance Scoring: Probabilistic ranking ensures the most relevant document matches appear at the top of results.
- Fuzzy Title Matching: Find notes instantly even with minor typos or incomplete title phrasing.
- Lightweight System Footprint: The search index consumes minimal RAM and updates asynchronously in background threads.
Lightning-Fast Query Execution at Any Scale
The embedded SQLite FTS5 index scans millions of words in single-digit milliseconds, ensuring that knowledge retrieval is instantaneous even as your archive expands to tens of thousands of notes over years of use.
BM25 scoring ranks exact title matches above body text matches, so typing a few letters in the Quick Open palette brings up the exact note you need before you even finish typing.
