Livecoding an SMS Chatbot
summary
Live-codes an SMS/MMS chatbot that impersonates a Twitter celebrity using markov chains, pulling up to 3,000 historical tweets via the Twitter gem (200 per request), cleaning out links/RTs/hashtags, and seeding a Markov model (marky_markov) saved to CSV. Uses Sinatra + Twilio to reply to incoming texts with generated lines and send an image, and uses the Twilio REST API to place calls that play a repeated robot-voiced Markov sentence — demo works live and returns plausible, often funny, DHH-like output.
AI-written summary · about
chapters (6)
- 0:00Introduction to bot building
- 1:03Understanding Markov chains
- 2:20Scraping and training data
- 6:36Building the SMS bot
- 11:31Implementing voice bot calls
- 14:40Advanced tools and closing








