Skip to content

First Steps

pixelnull edited this page Mar 25, 2026 · 5 revisions

First Steps

You've installed DeepLore Enhanced and connected your vault (Quick Start). Now what? This guide walks you through building a useful lore vault and tuning the extension for your needs.

Understanding the Pipeline

DeepLore works in three modes:

Mode How It Works Cost Best For
Keyword Only Scans recent chat messages for keywords defined in your entries Free Small vaults (<50 entries), simple setups
Two-Stage Keywords pre-filter candidates, then AI picks the best matches 1-2 API calls/generation Medium vaults (50-200 entries)
AI Only AI evaluates the full vault directly 1+ API calls/generation Large vaults, complex lore, best results

Start with Keyword Only to learn the system, then upgrade to Two-Stage when your vault grows.

Writing Your First Entries

A good starter vault has 3-5 entries covering:

1. A Character Entry

---
tags: [lorebook]
keys: [Eris, the Spymaster]
priority: 20
summary: "Eris, the protagonist's spymaster and closest enforcer. Select when espionage, intelligence, interrogation, or loyalty comes up."
---
  • Priority 20 = important character (inner circle)
  • Keys include name and title/role

2. A Location Entry

---
tags: [lorebook]
keys: [The Crimson Den, tavern, underground bar]
priority: 50
summary: "An underground tavern serving as a neutral meeting ground. Select when taverns, drinking, underground dealings, or nightlife come up."
---
  • Priority 50 = standard importance
  • Keys include the name AND generic terms that would trigger it

3. A Lore Concept Entry

---
tags: [lorebook]
keys: [bloodbond, blood bond, feeding dependency]
priority: 35
summary: "The biological dependency created when a vampire feeds from a mortal. Select when feeding, biting, addiction, or chattel dynamics come up."
---
  • Priority 35 = core lore
  • Multiple keyword variants catch different phrasings

Constants vs Triggered Entries

  • Regular entries (just lorebook tag): Only injected when keywords match
  • Constants (add lorebook-always tag): Always injected, every generation. Use for core world rules, character sheets, or writing instructions
  • Bootstrap entries (add lorebook-bootstrap tag): Force-injected when chat is short (first few messages), then become regular entries
  • Seed entries (add lorebook-seed tag): Content sent to AI as context for better entry selection on new chats (NOT injected into the writing AI)

Rule of thumb: Keep constants minimal. Every constant uses context budget on every generation.

Tuning Scan Depth and Budget

Scan Depth

How many recent chat messages DeepLore scans for keywords:

  • 2-3: Only matches recent conversation topics (responsive but narrow)
  • 4-6: Good default balance (recommended starting point)
  • 8-10: Broader matching, may pull in less relevant entries
  • 0: Disables keyword scanning entirely (use with AI-only mode)

Token Budget

Maximum tokens DeepLore can inject per generation:

  • 1000-2000: Conservative, good for smaller contexts
  • 3000-5000: Balanced, works for most setups
  • Unlimited: Let DeepLore use as much as it needs (watch your context usage)

Start with scan depth 4 and budget 3000, then adjust based on /dle-inspect results.

Enabling AI Search

When you're ready for smarter matching:

  1. Go to Search Mode → select Two-Stage
  2. In the AI Search section of settings, pick a connection mode:
    • Profile: Uses a SillyTavern Connection Manager profile (recommended)
    • Proxy: Routes through a CORS proxy (for advanced setups)
  3. Select or create a connection profile with a fast, cheap model (Claude Haiku, GPT-4o-mini, etc.)
  4. Test with /dle-status — it should show "AI search: enabled"

AI search shines when your vault is large or entries have nuanced triggers that keywords can't capture.

Checking Entry Quality

Run /dle-health regularly. It checks for:

  • Missing or empty keywords
  • Duplicate keywords across entries
  • Missing summaries (needed for AI search)
  • Entries too large (may dominate your budget)
  • Broken wikilinks
  • Configuration issues

Aim for grade A or A+.

Verifying What's Injecting

Three tools help you see what's happening:

  1. /dle-inspect — Shows the last pipeline trace: what matched, why, and what was injected
  2. Context Cartographer (book icon on AI messages) — Shows sources from the last generation with token counts
  3. /dle-simulate — Replays the entire chat showing which entries activate at each message

Common Early Mistakes

Mistake Fix
Too many constants Keep constants under 5-10. Use keywords for everything else
Keywords too generic "the", "and", "said" will match everything. Use specific terms
Keywords too specific "Lord Vexathorn the Undying" won't match "Vexathorn". Add aliases
No summaries AI search needs summaries to work. Write them for every entry
Huge entries (5000+ tokens) Split into smaller focused entries or use scan depth overrides
All priority 50 Use the full range: 20 (critical), 35 (core), 50 (standard), 60+ (flavor)

Next Steps

Clone this wiki locally