Quick Start
Key Features
| Feature | Description |
|---|---|
| Trigram Search | Fuzzy file/content search (~100ms) |
| Call Chain | Callers/callees tracing with BFS (7-496ms) |
| Tree View | Hierarchical DFS call tree |
| Cross-Convention | createSession finds create_session automatically |
| Multi-Language | TypeScript, Rust, Python, Go, Zig |
| Tauri IPC Bridge | Links TS invoke('cmd') to Rust #[tauri::command] fn cmd |
| Closure Tracking | Tracks calls inside callbacks and spawn blocks |
Supported Languages
| Language | Extensions | Symbol Types |
|---|---|---|
| Rust | .rs | fn, struct, enum, trait, impl |
| TypeScript | .ts, .tsx | function, class, interface, type |
| JavaScript | .js, .jsx | function, class |
| Python | .py | def, class |
| Go | .go | func, type, struct |
Resolution Modes
Always try--strict first, fall back to fuzzy if needed.
| Mode | When to Use |
|---|---|
--strict | Start here. Precise results. |
| Default | When --strict returns too few results |
--fuzzy=0.7 | Relaxed fuzzy, fewer false positives |
--fuzzy=0.4 | Aggressive fuzzy, more matches |
Index Management
-r flag forces a full reindex before the command. Unconditional — no stale data.
| Mode | Time | Notes |
|---|---|---|
Without -r | 50-200ms | Fast, may be outdated |
With -r | 2-5s | Slow, guaranteed fresh |
Behavioral Engineering
Monogram is not just a search tool — it’s a behavioral engineering system that reduces LLM uncertainty:- Avoidance reduction: 70% → 20%
- Completion quality: 40% → 90%
- Confidence increase: 30% → 85%