Skip to main content

Dwarf Fortress storytelling assistant — turns game events into narrative stories

Project description

df-storyteller

A fortress journal and storytelling companion for Dwarf Fortress. Captures game events, dwarf personalities, and world history through DFHack. Use AI to generate narratives, or write your own — the choice is yours.

Python 3.11+ License

Full Documentation on the Wiki

Features

Two Modes: AI or Manual

  • No-LLM Mode — Use as a structured fortress journal. Write your own chronicles, biographies, gazette editions, sagas, and quests. No AI required.
  • LLM Mode — AI generates narratives grounded in your gameplay. Manual writing is still available alongside AI output.
  • Setup asks upfront — choose your preference during init, change anytime in Settings.

Narrative Generation (AI or Player-Written)

  • Fortress Chronicles — Seasonal narratives tracking what's changing in your fortress. Edit entries inline.
  • Dwarf Diaries — First-person journal entries shaped by personality, beliefs, and stress
  • Character Biographies — Dated entries that evolve as dwarves change over time
  • Death Eulogies — Memorial narratives for fallen dwarves
  • Battle Reports — Dramatic combat accounts written by survivors or the fortress chronicler
  • Epic Sagas — World history narratives from legends data
  • Fortress Gazette — A dwarven newspaper with five sections. Write your own or generate with AI. Edit published editions.

Player Agency

  • Dwarf Highlights — Mark dwarves as Protagonist, Antagonist, or Watchlist. Highlighted dwarves get more focus in AI narratives and show role badges across the UI.
  • Player Notes — 8 tag types (Suspicion, Fact, Theory, Rumor, Secret, Foreshadow, Mood, What If) that influence how AI writes.
  • Manual Writing Everywhere — Write your own entries on every page, even with AI enabled.
  • Inline Editing — Edit chronicles, quests, and gazette editions after creation.

Quest System

  • AI-Generated or Player-Created — Generate quests from fortress state, or create your own with category and difficulty
  • Resolve with Comments — Complete quests with a player-written resolution (no AI needed)
  • Edit on the Go — Update quest titles and descriptions inline
  • Difficulty Tiers — Easy, Medium, Hard, Legendary
  • Completion Narratives — AI writes how the quest was fulfilled (LLM mode), or resolve manually

Visualization & Data

  • Relationship Web — Interactive force-directed graph of fortress connections
  • Pantheon — Deity worship chart with sphere descriptions from legends
  • Combat Log — Blow-by-blow fight details with siege grouping
  • Chat Log — Dwarf conversations with AI social summaries
  • Lore Browser — Searchable world history with hover tooltips (kill counts, battle forces, relationships)
  • Live Event Feed — Real-time game events via WebSocket

Screenshots

Chronicle

AI-generated seasonal narratives that track migrations, conflicts, role changes, and the evolving story of your fortress. Includes fortress-wide player notes with 8 tag types that influence how the AI writes. Chronicle

Dwarves

Character sheets with tabbed Notes, Biography, and Diary sections. Personality traits, skills, relationships, combat record, and equipment. First-person diary entries written in the dwarf's voice based on their personality. Dwarves

Relationship Web

Interactive force-directed graph showing family, friend, rival, and deity connections across the entire fortress. Drag, zoom, hover for details, double-click to focus on a dwarf's connections. Relationship Web

Pantheon

Bar chart of deity worship across the fortress, sorted by worshipper count. Click any deity to see who worships them. Sphere descriptions (death, minerals, sacrifice, etc.) pulled from legends data. Pantheon

Events

Live event feed via WebSocket, collapsible combat log with blow-by-blow fight details and siege grouping, battle reports written by survivors, and a chat log of dwarf conversations with AI social summaries. Events

Lore

Searchable world history browser covering civilizations, wars, battles, historical figures, artifacts, and more. Hover tooltips show rich detail — kill counts, battle forces, relationships, deity spheres, artifact descriptions. Lore

Quests

AI-generated quests based on your actual fortress state — citizens, buildings, events, religion, military. Filtered by category and difficulty. Grounded in real DF mechanics. Completion narratives feed into future chronicles. Quests

Gazette

A dwarven newspaper with five sections: The Fortress Herald, Military Dispatches, Quarry Gossip, Quest Board, and Obituaries. Written by the fortress's best writer in their personality voice. Newspaper-style two-column layout. Gazette

Quick Start

Prerequisites: Python 3.11+ — during install, check "Add Python to PATH".

Install and run:

pip install df-storyteller
python -m df_storyteller init
python -m df_storyteller serve

In DFHack console (first time per fortress):

storyteller-begin

Or install from source:

git clone https://github.com/Been012/df-storyteller.git
cd df-storyteller
pip install -e .
python -m df_storyteller init
python -m df_storyteller serve

Requirements

  • Dwarf Fortress v50.x (Steam / DF Premium) — tested with v50.14
  • DFHack v50.14-r1+ (Steam Workshop or dfhack.org)
  • Python 3.11+
  • An LLM provider (optional — not needed for No-LLM mode):
    • Ollama — free, runs locally, no API key needed (supports thinking models)
    • Anthropic Claude — API key required
    • OpenAI — API key required

Note: This tool uses DF Premium (Steam) APIs. Classic DF (pre-Steam) is not supported — some DFHack fields like unit.relations don't exist in the Steam version. We use histfig_links on historical figures instead.

Tech Stack

  • Backend: Python 3.11+, FastAPI, Pydantic v2
  • Frontend: Jinja2 templates, vanilla CSS/JS (no build step)
  • Game integration: DFHack Lua scripts
  • LLM: Anthropic SDK, OpenAI SDK, Ollama REST API (with thinking model support)
  • All narratives grounded in DF mechanics — the AI knows squad sizes, siege thresholds, temple values, what players can and cannot control

Documentation

See the Wiki for:

Notes

  • This tool is designed to run locally on your machine (localhost). It is not intended to be exposed to the internet or run on a public server.
  • Developed and tested on Windows with DF Premium (Steam). Should work on Mac and Linux — all code uses cross-platform libraries (pathlib, FastAPI) and DFHack APIs are platform-independent — but these platforms are untested. If you run into issues, please open a bug report.
  • Config and stories are stored at ~/.df-storyteller/. API keys are in config.toml.

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

df_storyteller-0.3.0.tar.gz (5.1 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

df_storyteller-0.3.0-py3-none-any.whl (173.0 kB view details)

Uploaded Python 3

File details

Details for the file df_storyteller-0.3.0.tar.gz.

File metadata

  • Download URL: df_storyteller-0.3.0.tar.gz
  • Upload date:
  • Size: 5.1 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for df_storyteller-0.3.0.tar.gz
Algorithm Hash digest
SHA256 2363689c33f92b635d3b327f5b4ff8343f3a7016996c73d6bc3ded4e336c3202
MD5 3a8d0992216c1cf44ccd956945add8e1
BLAKE2b-256 d066edb253f370e098aadd74c5d09f6107768a6abc9a3e6410f76ebd6adb44a2

See more details on using hashes here.

File details

Details for the file df_storyteller-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: df_storyteller-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 173.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for df_storyteller-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 0cd9bfab4cada021013c4036335131582a42d4e75c27862a4f34c4d08591744a
MD5 c2f567a46b7bfd0aacb7a2e9bb2b140c
BLAKE2b-256 2fa0d6ac4f036b5b6268683589e9396fd0e4f50d553ba8de0b21bd80ebc0ab2c

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page