Settings
Complete reference for all BlazeTracker settings.
Connection
| Setting | Type | Default | Description |
|---|---|---|---|
| Connection Profile | select | — | API connection for extraction calls. See Setup. |
| Auto Extract | boolean | true | Automatically extract state from new messages. |
Display
| Setting | Type | Default | Description |
|---|---|---|---|
| State Display Position | 'above' | 'below' | 'below' | Show state block above or below message content. |
| Temperature Unit | 'fahrenheit' | 'celsius' | 'fahrenheit' | Display unit for temperatures. |
| Time Format | '12h' | '24h' | '12h' | Time display format. |
Tracking
Enable or disable extraction modules. Disabling modules reduces LLM calls per message. See Track Dependencies for dependency rules.
| Module | Default | What It Tracks |
|---|---|---|
| Time | true | Narrative date and time |
| Location | true | Area, place, position |
| Props | true | Nearby objects and items |
| Climate | true | Weather and temperature (procedural) |
| Characters | true | Positions, activities, moods, outfits |
| Relationships | true | Feelings, secrets, wants, status |
| Scene | true | Topic, tone, tension |
| Narrative | true | Events, milestones, chapters |
Message Limits
| Setting | Type | Default | Description |
|---|---|---|---|
| Max Messages to Send | number | 10 | Maximum recent messages included in extractor prompts. |
| Max Chapter Messages | number | 24 | Maximum messages sent to the chapter description extractor. |
Injection
| Setting | Type | Default | Description |
|---|---|---|---|
| Auto Inject State | boolean | true | Automatically inject scene state (time, location, characters, etc.) into prompts. Disable for macro-only workflows. |
| Auto Inject Narrative | boolean | true | Automatically inject chapter summaries and events into prompts. Disable for macro-only workflows. |
| Injection Depth | number | 0 | Prompt injection depth (0 = near most recent messages). |
| Max Recent Chapters | number | 5 | Maximum past chapters in “Story So Far” injection. |
| Max Recent Events | number | 15 | Maximum out-of-context events from current chapter. |
| Token Budget | number | 0 | Token budget for injection (0 = use ST’s context size). |
Scene Shakeups
| Setting | Type | Default | Description |
|---|---|---|---|
| Enable Scene Shakeups | boolean | false | Inject random context-appropriate events into prompts. See Scene Shakeups guide. |
| Max Messages Between Shakeups | number | 20 | Messages at which trigger probability reaches 100%. Range: 5–100. |
Advanced
LLM Settings
| Setting | Type | Default | Description |
|---|---|---|---|
| Max Tokens | number | 500 | Maximum tokens for LLM responses. |
| Max Requests/Min | number | 0 | Rate limit for LLM requests (0 = no limit). |
| Include World Info | boolean | false | Include lorebook entries in extractor prompts. |
| Prompt Prefix | string | '' | Prepended to user template of all prompts (e.g., /nothink). |
| Prompt Suffix | string | '' | Appended to user template of all prompts. |
Category Temperatures
Default LLM temperatures per extraction category. Individual prompts can override these.
| Category | Default | Notes |
|---|---|---|
| Time | 0.3 | Low for deterministic time parsing |
| Location | 0.5 | Moderate |
| Props | 0.5 | Moderate |
| Climate | 0.3 | Low for consistent weather classification |
| Characters | 0.5 | Moderate |
| Relationships | 0.6 | Slightly higher for nuanced feelings |
| Scene | 0.5 | Moderate |
| Narrative | 0.6 | Slightly higher for creative summaries |
Debug
| Setting | Type | Default | Description |
|---|---|---|---|
| Debug Logging | boolean | false | Log debug information to browser console. |
Custom Prompts
Each extraction prompt can be overridden individually. For each prompt:
| Setting | Description |
|---|---|
| Temperature | LLM temperature for this specific prompt (overrides category default) |
| System Prompt | Static instructions (cacheable) |
| User Template | Dynamic content with placeholders |
See Custom Prompts guide for details on customization.