Skip to content

Features

Jump to bottom
pixelnull edited this page Mar 21, 2026 · 15 revisions

Features

This page covers all of DeepLore Enhanced's features in detail. For the core matching pipeline, see Pipeline. For AI Search specifics, see AI Search.

Context Cartographer

Adds a book icon button to each AI message's action bar. Click it to see which vault entries were injected for that message, why they matched, and how much context they used.

The popup shows:

  • Entry name (clickable link to Obsidian if vault connection names match Obsidian vault names)
  • Match type: keyword, AI (with confidence and reason), constant, pinned, or bootstrap
  • Priority value
  • Token cost with color-gradient bar (green/yellow/red relative to vault average token size)
  • Entries grouped by injection position (Before Main Prompt, In-chat @depth, After Main Prompt)
  • Generation-to-generation diff: shows +new and -removed entries with reasons (e.g., "Bootstrap fall-off", "No longer matched")
  • Expandable content preview (click to show first 300 chars with keyword highlighting)
  • Vault source label (when multiple vaults are connected)
  • Entry metadata: keys, requires, era, location, and wikilinks

Setup:

  1. Enable "Show Lore Sources Button" in Settings Reference
  2. Set vault connection names to match your Obsidian vault names exactly to enable deep links

Notes:

  • Source data is saved per-message in message.extra.deeplore_sources, so it persists across sessions
  • The book icon only appears on messages that have lore source data

Author's Notebook

A persistent per-chat scratchpad that is injected into every generation. Use it for author notes, scene direction, tone guidance, or anything you want the writing AI to always see for this specific chat.

How it works:

  1. Open the notebook via /dle-notebook or the "Open Notebook" button in settings
  2. Write any text — it's saved per-chat in chat_metadata
  3. The notebook content is injected into every generation as a separate prompt, independent of the entry pipeline

Setup:

  1. Enable "Enable Author's Notebook" in Settings Reference
  2. Choose injection position (Before Main Prompt, After Main Prompt, or In-chat @ Depth)
  3. Open the editor and start writing

Notes:

  • Notebook content is stored in chat_metadata.deeplore_notebook — it persists across sessions for that chat
  • Injection is independent of the lorebook pipeline — the notebook always injects when enabled, regardless of matching
  • Has its own position, depth, and role settings separate from the main lorebook injection

Auto Lorebook

AI analyzes your chat for characters, locations, items, and concepts that are mentioned but don't have an existing lorebook entry, then suggests new entries you can review and accept.

How it works:

  1. After every N messages (configurable), or on-demand via /dle-newlore, the AI scans recent chat
  2. It compares against existing entries and identifies gaps
  3. Suggestions appear in a popup with title, type, keywords, summary, and content
  4. Accept to write the entry to Obsidian, or reject to skip

Connection options:

  • SillyTavern (default): Uses your active AI connection
  • Connection Profile: Use any saved Connection Manager profile
  • Custom Proxy: Use a separate proxy server

Setup:

  1. Enable "Enable Auto Lorebook" in Settings Reference
  2. Set the trigger interval (every N messages)
  3. Optionally set a target folder for new entries
  4. Choose a connection mode

Notes:

  • Existing entries are filtered out (case-insensitive title match)
  • Accepted entries are written with proper frontmatter (type, priority, tags, keys, summary)
  • Use /dle-newlore to trigger on-demand without enabling automatic suggestions

Multi-Vault Support

Connect multiple Obsidian vaults simultaneously. Each vault has its own port and API key. Entries from all enabled vaults are merged into a single index.

Setup:

  1. In the "Vault Connections" section, click "Add Vault" to add multiple vault connections
  2. Each vault has a name, port, API key, and enable toggle
  3. Click "Test All" to verify all connections

Notes:

  • Entries from all enabled vaults are merged and treated identically
  • Each entry tracks its vaultSource for diagnostics
  • The health check validates multi-vault configuration

Session Scribe

Automatically summarizes your roleplay sessions and writes them as timestamped markdown notes to your Obsidian vault.

How it works:

  1. Tracks actual chat position — after every N new messages (configurable), Scribe generates a summary
  2. The summary is written as a markdown file to the configured Session Folder in your vault
  3. Notes include frontmatter with timestamp, character name, and chat ID
  4. Each summary builds on the previous one — the prior note is fed as context so the AI doesn't repeat itself

On-demand: Use /dle-scribe to write a summary at any time. Optionally pass a focus topic: /dle-scribe What happened during the trial?

Connection options:

  • SillyTavern (default): Uses your active AI connection via generateQuietPrompt
  • Connection Profile: Use any saved Connection Manager profile — lets you route summaries through a different model/provider
  • Custom Proxy: Use a separate proxy server (claude-code-proxy or compatible Anthropic Messages API endpoint)

Setup:

  1. Enable "Enable Session Scribe" in Settings Reference
  2. Set the auto-scribe interval (every N messages)
  3. Set the Session Folder (default: Sessions)
  4. Choose a connection mode (SillyTavern, Connection Profile, or Custom Proxy)
  5. Optionally customize the summary prompt and message window depth

Notes:

  • Writes directly to the Obsidian vault via the Local REST API
  • Default prompt covers events, character dynamics, revelations, and state changes in past tense
  • Configurable message window (default: 20 messages) and response token limit (default: 1024)

Vault Change Detection & Auto-Sync

When the index rebuilds, DeepLore compares the new index against the previous one and reports changes.

Detects:

  • New entries added
  • Entries removed
  • Modified content
  • Changed keywords

Auto-Sync Polling: Set "Auto-Sync Interval" to automatically re-check the vault every N seconds. When changes are detected, toast notifications summarize what changed (if "Show Sync Change Toasts" is enabled).

Manual refresh: Click "Refresh Index" in settings or use /dle-refresh.

Cooldown Tags

Per-entry cooldown: N in frontmatter. After an entry triggers, it's skipped for the next N generations before becoming eligible again.

Use case: Prevent the same lore from being re-injected every single generation. Useful for flavor text or background entries that don't need constant repetition.

Example:

cooldown: 3  # After triggering, skip for 3 generations

Notes:

  • Cooldown is tracked per-session (resets on chat change or page refresh)
  • Constants (#lorebook-always) are exempt from cooldown

Warmup Tags

Per-entry warmup: N in frontmatter. An entry's keywords must appear N or more times in the scan text before it triggers for the first time.

Use case: Prevent entries from triggering on a single casual mention. Ensure a topic is being discussed in depth before injecting detailed lore.

Example:

warmup: 3  # Keyword must appear 3+ times in scan text before first trigger

Notes:

  • The warmup threshold is checked every generation — the keyword must meet the hit count each time, not just the first time
  • Count is based on occurrences in the scan text, not unique messages

Re-injection Cooldown

Global setting (not per-entry). Skips re-injecting an entry for N generations after it was last injected. Helps save context by avoiding redundant lore repetition.

Setup: Set "Re-injection Cooldown" in Settings Reference (0 = disabled)

Notes:

  • Constants (#lorebook-always) are exempt
  • Tracked per-session (resets on chat change)

Injection Deduplication

Global setting that prevents the same entries from being injected in consecutive generations. When enabled, entries that were injected within the last N generations (configurable lookback depth) are skipped.

Setup:

  1. Check "Strip Duplicate Injections" in Settings Reference
  2. Set the "Lookback Depth" (default 2 — checks last 2 generations)

Notes:

  • Constants (#lorebook-always) are exempt — they always inject
  • Injection history is tracked per-chat in chat_metadata.deeplore_injection_log
  • Different from Re-injection Cooldown: deduplication checks a sliding window of recent generations, while re-injection cooldown counts generations since last injection

Active Character Boost

When enabled, automatically matches the active character's vault entry by name or keyword, even if the character isn't mentioned in recent chat messages.

Use case: Ensure the character you're roleplaying with always has their lore available, without relying on their name appearing in the most recent messages.

Setup: Check "Active Character Boost" in Settings Reference

Conditional Gating

Entries can declare dependencies on other entries using requires and excludes frontmatter fields.

requires

All listed entry titles must be in the matched set for this entry to activate.

requires:
  - Eris
  - Dark Council

This entry only injects when both "Eris" and "Dark Council" are also matched.

excludes

If any listed entry title is in the matched set, this entry is blocked.

excludes:
  - Draft Notes

This entry is blocked if "Draft Notes" is matched.

Cascading Resolution

Gating resolves iteratively. If Entry A requires Entry B, and Entry B gets removed by its own gating rules, Entry A is also removed. This cascading continues until no more entries are affected.

See Writing Vault Entries for a complete template.

Refine Keys

Per-entry refine_keys in frontmatter. Adds a secondary AND filter on top of primary keyword matching. When set, at least one refine key must also appear in the scan text for the entry to trigger.

Use case: Reduce false positives for entries with common primary keywords. For example, a character named "Rose" might have refine_keys requiring mention of their faction or role to avoid triggering on every use of the word "rose."

Example:

keys:
  - Rose
  - Rose Blackwood
refine_keys:
  - guild
  - spymaster
  - intelligence

See Writing Vault Entries for a complete template.

Cascade Links

Per-entry cascade_links in frontmatter. When an entry matches, all entries listed in its cascade_links are automatically pulled in -- no keyword check needed for the linked entries.

Use case: Ensure related entries always travel together. Unlike wikilink recursion (which requires keyword matches), cascade links are unconditional.

Example:

cascade_links: ["Soulbrand Removal", "Ironveil Guild"]

When this entry matches, "Soulbrand Removal" and "Ironveil Guild" are automatically included.

See Writing Vault Entries for a complete template.

Per-Entry Injection Position

Entries can override the global injection position via frontmatter:

Field Values Description
position before, after, in_chat Where to inject
depth number Chat depth (for in_chat)
role system, user, assistant Message role (for in_chat)

Entries are grouped by their effective position (global default or override) and each group is injected separately.

Example: You might want most lore injected at depth 4 as system messages, but a character's dialogue hints injected at depth 1 as user messages.

See Writing Vault Entries for templates.

Prompt Manager Integration

Set Injection Mode to Prompt List to register DLE's injections as named entries in SillyTavern's Prompt Manager. This lets you drag them to any position in the prompt order — before character definition, after Author's Note, between example messages, wherever.

How it works:

  1. Switch injection mode to "Prompt List" in DLE settings
  2. Generate at least once so the entries appear
  3. Open the Prompt Manager and find deeplore_constants and deeplore_lore
  4. Drag them to your desired position, or switch to Absolute mode with a custom depth

Notes:

  • Requires a Chat Completion API (OpenAI-compatible)
  • Per-entry frontmatter overrides with custom position/depth still create separate injection groups
  • The deeplore_notebook entry also appears in the PM (it already uses a stable key)

New Chat Features

On a brand new chat (below the New Chat Threshold, default 3 messages), two features help bootstrap the conversation:

Seed Entries (#lorebook-seed)

  • Entry content is sent to the AI as additional story context alongside the chat
  • Helps the AI understand your setting and make better entry selections from minimal context
  • NOT injected into the writing AI's context. Only informs AI search.
  • When seed mode is active, AI is instructed to fill to maxEntries selections (more aggressive)

Bootstrap Entries (#lorebook-bootstrap)

  • Force-injected like constants when chat is short
  • Once chat grows past the threshold, they become regular entries managed by normal selection
  • Good for writing instructions or foundational lore needed at the start

An entry can have both tags: its content feeds the AI (seed) AND it force-injects (bootstrap).

Entry Analytics

Track how often each entry is matched and injected across generations. View with /dle-analytics.

Shows:

  • Table sorted by injection count
  • "Never Injected" section for dead entry detection

Use case: Identify entries with bad keywords that never trigger, or entries that trigger too frequently.

Entry Health Check

Audit all vault entries for common issues with /dle-health. Runs 30+ checks across multiple categories.

Check categories:

  • Multi-vault: Enabled vaults, API key validation
  • Settings: Scan depth disabled, AI mode without profile, proxy URL missing, budget too low, cache TTL, index staleness
  • Entry config: Duplicate titles, empty keys, empty content, orphaned requires/excludes/cascade_links, requires + excludes same title, oversized entries (>1500 tokens), missing summary
  • Gating: Circular requires, unresolved links, conflicting overrides
  • AI Search: Entries without summary fields
  • Keywords: Short keywords (2 chars or less), duplicate keywords across entries
  • Size: Constants exceeding budget, seed entries >2000 tokens
  • Injection: Depth/role override without in_chat position
  • Entry behavior: Cooldown on constants, warmup unlikely to trigger, bootstrap with no keywords, probability zero, excluded from recursion with no keywords

Entry Browser

Browse all indexed entries in a searchable, filterable popup with /dle-browse or the "Browse" button in settings. Useful for quickly reviewing your vault without switching to Obsidian.

Relationship Graph

Visualize entry relationships as an interactive force-directed graph with /dle-graph. Shows wikilinks, requires/excludes connections, and cascade links between entries.

Simulation

Replay your chat history step-by-step with /dle-simulate, showing which entries activate and deactivate at each message. Useful for understanding how your keywords and pipeline behave across a conversation.

Optimize Keys

Use /dle-optimize-keys <entry name> to have AI analyze an entry and suggest better keywords. The AI considers the entry's content, summary, and current keywords to recommend improvements.

Pipeline Inspector

View a detailed trace of the last generation with /dle-inspect.

Shows:

  • Pipeline mode (two-stage, ai-only, keywords-only)
  • Keyword matches with trigger keywords
  • AI selections with confidence and reasons
  • Fallback status
  • Constants and bootstrap entries

See Slash Commands for all available commands.

Per-Chat Pin/Block

Pin entries to always inject or block entries from injecting, on a per-chat basis. Pins and blocks are stored in chat_metadata and survive page reloads.

Commands:

  • /dle-pin <entry name> — Pin an entry (always inject in this chat)
  • /dle-unpin <entry name> — Remove a pin
  • /dle-block <entry name> — Block an entry (never inject in this chat)
  • /dle-unblock <entry name> — Remove a block
  • /dle-pins — Show all pins and blocks for the current chat

How it works:

  • Pinned entries are force-injected like constants, regardless of keywords or AI selection
  • Blocked entries are removed from the pipeline before injection, regardless of matches
  • Pins/blocks apply after the main pipeline runs but before formatting

Contextual Gating

Filter entries based on the current story context using frontmatter fields: era, location, scene_type, and character_present. Set the active context with slash commands, and entries that don't match the current context are filtered out.

Frontmatter fields:

Field Type Description
era string Entry only injects when the active era matches
location string Entry only injects when the active location matches
scene_type string Entry only injects when the active scene type matches
character_present string[] Entry only injects when any listed character is present

Commands:

  • /dle-set-era [era] — Set the active era. With no argument, opens a browse-and-select popup showing all era values in your vault with entry counts
  • /dle-set-location [location] — Set the active location. With no argument, shows a browse-and-select popup
  • /dle-set-scene [type] — Set the active scene type. With no argument, shows a browse-and-select popup
  • /dle-set-characters <names> — Set present characters (comma-separated)
  • /dle-context-state — Show the current contextual gating state

Notes:

  • Context state is stored per-chat in chat_metadata.deeplore_context
  • Entries without contextual fields are unaffected (always pass through)
  • Partial matches work: an entry with only era set is filtered only on era, regardless of location/scene/character

Entry Decay & Freshness

Tracks how many generations have passed since each entry was last injected. Stale entries (not seen recently) get a boost in the AI manifest; frequently injected entries get a penalty. This naturally rotates lore without manual intervention.

Setup:

  1. Enable "Entry Decay" in Settings Reference
  2. Set the Boost Threshold (default 5) — generations without injection before freshness boost
  3. Set the Penalty Threshold (default 2) — consecutive injections before frequency penalty

Notes:

  • Decay tracking is per-session (resets on chat change or page refresh)
  • Only affects AI search manifest (adds decay hints for the AI to consider)
  • Constants are exempt from decay penalties

ST Lorebook Import Bridge

Convert SillyTavern World Info JSON exports into Obsidian vault notes with proper frontmatter. Handles three formats: WI export JSON, V2 character cards with embedded WI, and raw entry arrays.

Usage: /dle-import opens a popup where you paste your WI JSON and choose a target folder.

What it converts:

  • key → keys (primary keywords)
  • keysecondary → refine_keys
  • order → priority
  • position → position (mapped to before/after/in_chat)
  • depth → depth
  • probability → probability (scaled from 0-100 to 0.0-1.0)
  • constant → #lorebook-always tag
  • comment → entry title

Auto-Summary Generation

Generate AI summaries for entries that don't have a summary field. Good summaries improve AI search quality significantly.

Usage: /dle-summarize scans all indexed entries, identifies those without summaries, and generates AI summaries one at a time. Each summary is written directly to the entry's frontmatter in Obsidian.

Setup Wizard

A guided first-time setup experience. Walks through Obsidian vault connection, tag configuration, and search mode selection. AI search connection (profile or proxy) must be configured separately in the settings panel.

Usage: /dle-setup or the Setup button in the Quick Actions bar.

Quick Actions Bar

A toolbar of one-click buttons at the top of the settings panel for common operations. Includes two rows: always-visible actions (Browse, Map, Health, Refresh) and an expandable "More" row (Graph, Simulate, Analytics, Optimize, Inspect, Setup).

Uses SillyTavern's standard menu_button menu_button_icon pattern with Font Awesome icons. All buttons call their functions directly (no slash command roundtrip).

Prompt Cache Optimization

In proxy mode, the AI search manifest is placed first in the message payload with cache_control breakpoints. This leverages Anthropic's prompt caching so that the manifest (which rarely changes between calls) is cached server-side, reducing token costs on subsequent calls.

Only applies to Custom Proxy mode. Connection Profile mode does not support cache_control breakpoints.

Circuit Breaker

The Obsidian REST API connection uses a circuit breaker pattern to avoid hammering a down server. States: closed (normal), open (failing — skip calls for backoff period), half-open (try one test call).

Exponential backoff from 2s to 15s. Automatic — no settings to configure. Resets when a call succeeds.

IndexedDB Persistent Cache

The parsed vault index is saved to IndexedDB (DeepLoreEnhanced database, vaultCache store) after every successful build. On page load, the extension hydrates from IndexedDB instantly (no Obsidian call needed), then validates against Obsidian in the background.

Benefits:

  • Near-instant startup — lore is available before the first generation
  • Works even if Obsidian is briefly unreachable on page load
  • Automatic — no settings to configure

Reuse Sync

When auto-sync triggers, the extension fetches all vault file contents but avoids redundant work:

  1. Fetches all file contents from Obsidian (local fetch is fast)
  2. Computes content hashes and compares against the existing index
  3. Reuses already-parsed entries for unchanged files (skips parse + tokenize)
  4. Re-parses only new or modified files
  5. Removes entries for deleted files
  6. Falls back to full rebuild if the reuse approach fails

The savings come from skipping the expensive parse/tokenize step for unchanged entries, not from reducing network calls.

Hierarchical Manifest Clustering

For large vaults (40+ selectable entries), the AI search uses a two-call approach:

  1. Group entries by category (extracted from tags/type fields)
  2. First AI call: select relevant categories from the full list
  3. Second AI call: select specific entries from within those categories

Safety valve: if the category filter removes more than 80% of entries, it falls back to the full manifest. Requires at least 4 distinct categories to activate.

Sliding Window AI Cache

AI search caches results with a sliding window strategy. The manifest and chat context are hashed separately. When only new chat messages are appended (vault unchanged):

  • If the new messages don't contain any entity names/keys from the vault, cached results are reused
  • If new messages reference vault entities, the cache is invalidated and a fresh AI call is made

This means most regenerations, swipes, and non-lore-relevant messages reuse cached results automatically.

Scribe-Informed Retrieval

When enabled, the Session Scribe's latest summary is fed into the AI search context as additional story background. This gives the AI search a broader understanding of the ongoing narrative, helping it select entries relevant to the overall story arc rather than just the most recent messages.

Setup: Enable "Scribe-Informed Retrieval" in Settings Reference.

Confidence-Gated Budget

AI search over-requests entries (2x the configured max), then sorts results by confidence tier (high → medium → low) before applying the budget cap. High-confidence picks are prioritized, ensuring that when budget is limited, the most relevant entries make the cut.

DeepLore Wiki

Getting started

Features

Reference

Help

Clone this wiki locally