Winners of Inventiva 2024-1 on the "Best Performance" category!
Thanks to everybody, you're the real heroes :)
We are already developing a new version on Go
, with most of the functions migrated into Javascript
, give it a look: DnD-AI 2.0
Each year, 10 million Dungeons and Dragons sessions are created, but did you know that more than half of these cannot take place due to a lack of a Dungeon Master?
Dungeons and Dragons is a game that relies entirely on the imagination of its players, and among them, the Dungeon Master is the one who creates the story
the players will follow, develops characters, their abilities, monsters, maps, and items – essentially, the heart of the game
. We depend heavily on them, and there isn't always someone willing to take on the role.
What if we replaced them with an artificial intelligence
that is available at all times? One that allows you and your friends to play a quality session whenever you want, generating the map and enemies
automatically. That's us, DnD-AI, and we want to offer you this new way of playing – much more focused on immediate gameplay and with a faster pace.
Unlike similar ideas, we have various visual enhancements, such as real-time map
visualization and the generation of images
for the epic scenes that, until now, you could only imagine.
Our wiki:
https://github.com/QuitoTactico/DnD-AI/wiki
Our backlog:
https://github.com/users/QuitoTactico/projects/1
pip install -r "requirements.txt"
python manage.py runserver
Or sudo python3 manage.py runserver 0.0.0.0:80
if you want your own public instance.
- The needed dependencies and libraries defined in requirements.txt. To get them, run
pip install -r "requirements.txt"
- For image generation is needed either a
OpenAI API key
or aHuggingFace API key
. - For action interpretation, storytelling and campaign's info providing, a
Google Gemini API key
is needed. OpenAI GPT option will be re-added in the future.
You can provide this keys either creating a file called api_keys.env
on the root dir of the project.
# api_keys.env
# on root dir of the project
openai_api_key =
hf_api_key =
gemini_api_key =
Or a file called API.py
on /DnD_AI/
(the application, not the project). At the same level of functions_AI.py
, the file that uses the api keys.
# API.py
# on same folder as /DnD_AI/functions_AI.py
openai_api_key = " "
hf_api_key = " "
gemini_api_key = " "
In any case, the game is still playable with some (or none) of the API keys
. While AI generated images being replaced with placeholder images and action inputs not being interpreted, but supporting the use of commands instead.
DnD.AI.-.Opera.2024-05-05.23-05-21.mp4
For the EAFIT course: Integrator project 1