--:--:-- UTC LINK ONLINE BACK TO TIMELINE
PROJECT DOSSIER
2019

BabbleBot

Co-creator — three-person Hack Week team

A Discord bot that generates predictive text with Markov chains (plus a tongue-in-cheek 'Boomer Mode') — built for Discord Hack Week 2019.

aidiscord PythonMarkov ChainsNLPDiscord API
PROBLEM
Discord Hack Week: build something fun and shippable in a week. The idea — a bot that generates predictive/imitative text from chat, including a 'Boomer Mode' that apes an older-generation texting style.
APPROACH
A Python Discord bot driving Markov-chain text generation over chat input — lightweight, fast to build, and good for a hackathon timebox. The classic Markov approach to NLP-style predictive text, wired into Discord's bot API.
EXECUTION

Babble Bot is a Discord Hack Week result. Worked with 2 other people to implement a basic Markov chain generation engine with some fun variations like Boomer Mode, which made the bot modify replies to look more like elderly Facebook replies. We didn't get anything, but it was a fun project, and wrapping that functionality inside of a bot was fun. I have kept the bot going for some years now; one server still uses it. It would scrape 10000s of messages for a specific user and generate a Markov chain dictionary, then cache that and allow users to generate sentences they, or the entire server, might plausibly say. Sometimes, it was even coherent.

OUTCOME
A working predictive-text Discord bot shipped for Discord Hack Week 2019, built with a team of three. An early take on chat-driven text generation — predating the later Discord-bot and language work (MoD, Echo).
LINKS
▸ Repo