candlekeep

Query local PDF and Markdown books with page-level citations.

2|Updated Nov 6, 2025
One-click install
npx skills add https://github.com/SaharCarmel/Sahar-claude-code-marketplace --skill candlekeep
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: candlekeep
Source: https://github.com/SaharCarmel/Sahar-claude-code-marketplace/tree/main/plugins/candlekeep/skills/candlekeep
Command: npx skills add https://github.com/SaharCarmel/Sahar-claude-code-marketplace --skill candlekeep

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires alembic, markdown-it-py, pdfminer.six, python-dotenv, rich, sqlalchemy, typer, typer-cli, uvicorn, watchfiles, python-multipart, pydantic, pydantic-settings, fastapi, httpx, python-slugify, beautifulsoup4, lxml, html2text, pygments, mistune, mdit-py-plugins, mdit-py-toc, mdit-py-footnote, mdit-py-tasklist, mdit-py-abbr, mdit-py-attrs, mdit-py-deflist, mdit-py-frontmatter, mdit-py-math, mdit-py-mermaid, mdit-py-sub, mdit-py-sup, mdit-py-mark, mdit-py-ins, mdit-py-kbd, mdit-py-container, mdit-py-emoji, mdit-py-highlightjs, mdit-py-latex, mdit-py-linkify, mdit-py-strikethrough, mdit-py-table, mdit-py-url, mdit-py-wikilink, mdit-py-yaml, mdit-py-toml, mdit-py-csv, mdit-py-json, mdit-py-xml, mdit-py-html, mdit-py-rst, mdit-py-gfm, mdit-py-commonmark, mdit-py-markdown, mdit-py-pandoc, mdit-py-recommonmark, mdit-py-sphinx, mdit-py-text, mdit-py-plain, mdit-py-raw, mdit-py-code, mdit-py-mathjax, mdit-py-katex, mdit-py-plantuml, mdit-py-graphviz, mdit-py-vega, mdit-py-vega-lite, mdit-py-ditaa, mdit-py-blockdiag, mdit-py-seqdiag, mdit-py-actdiag, mdit-py-nwdiag, mdit-py-rackdiag, mdit-py-packetdiag, mdit-py-boxdiag, mdit-py-dot, mdit-py-flowchart, mdit-py-sequence, mdit-py-state, mdit-py-gantt, mdit-py-class, mdit-py-er, mdit-py-journey, mdit-py-git, mdit-py-pie, mdit-py-xy, mdit-py-timeline, mdit-py-quadrant, mdit-py-sankey, mdit-py-mindmap, mdit-py-c4, mdit-py-bpmn, mdit-py-uml, mdit-py-activity, mdit-py-component, mdit-py-deployment, mdit-py-object, mdit-py-package, mdit-py-profile, mdit-py-usecase, mdit-py-interaction, mdit-py-communication, mdit-py-composite, mdit-py-internal, mdit-py-timing, mdit-py-statechart, mdit-py-objectdiagram, mdit-py-componentdiagram, mdit-py-deploymentdiagram, mdit-py-packagediagram, mdit-py-profilediagram, mdit-py-usecasediagram, mdit-py-interactiondiagram, mdit-py-communicationdiagram, mdit-py-compositediagram, mdit-py-internaldiagram, mdit-py-timingdiagram, mdit-py-statechartdiagram.

What problem does it solve?

D&D 5e rulebooks and adventure modules are extensive, making quick lookups challenging during gameplay or preparation. This Skill centralizes and makes your D&D library instantly queryable, saving you time and reducing cognitive load.

Core Features & Use Cases

  • Centralized D&D Library: Store all your D&D 5e rulebooks and adventure modules in one accessible knowledge base.
  • Intelligent Content Querying: Quickly retrieve specific pages, sections, or information from any stored book.
  • Rule & Lore Reference: Instantly look up monster stats, spell descriptions, magic item properties, or lore details.
  • Use Case: During a D&D session, if a player asks about a specific spell effect or a monster's vulnerability, use this skill to query the relevant rulebook in seconds, keeping the game flowing smoothly.

Quick Start

List all D&D books currently stored in CandleKeep.

Frequently Asked Questions about candlekeep

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a searchable D&D knowledge base from PDFs and markdown files?

Create a local knowledge base by uploading D&D rulebooks and modules as PDFs or markdown. The Skill parses content, extracts metadata and table of contents, indexes it with SQLAlchemy, and makes it queryable through a FastAPI interface for instant rule and lore lookups during gameplay.

Can I query specific spells, monsters, and magic items from my D&D library with page citations?

Yes. Query your stored rulebooks to retrieve spell descriptions, monster stats, magic item properties, and lore details with precise page markers and citations. Results are grounded in exact source material from your local library.

What file formats does a D&D knowledge base support for rulebooks and adventure modules?

The Skill processes PDFs via pdfminer.six and markdown files. It parses both formats, extracts tables of contents, handles metadata, and deduplicates content using file hashes to build a unified, queryable index across your entire D&D library.

How do I prevent duplicate rulebooks when adding books to my knowledge base?

The Skill computes file hashes for deduplication, automatically detecting and preventing duplicate PDFs or markdown files from being indexed multiple times in your knowledge base.

Can I use this knowledge base to ground AI agent responses in official D&D source material?

Yes. The Skill gives AI agents local-first access to your D&D library, enforcing grounded, citable answers backed by exact page references and table of contents navigation, keeping responses tied to official rulebooks and modules.

Do I need internet access or cloud storage to run a D&D knowledge base?

No. The Skill enforces local-first storage and indexing using SQLAlchemy and python-dotenv for configuration. All PDFs and markdown files are processed and indexed locally without cloud dependencies, keeping your library private and offline-accessible.