Inspiration
As passionate Formula 1 fans and AI enthusiasts, we’ve always been captivated by the energy of live race commentary. But what if we could bring that same excitement into dashboards, race data apps, or even personal race reviews- all powered by AI? We wanted to build something that could narrate the thrill of the race not just with numbers, but with natural, context-aware commentary. That’s how BoxBoxBot was born- a voice from the pit wall, powered by data and driven by language.
What it does
BoxBoxBot transforms raw Formula 1 telemetry and timing data into real-time, driver-specific commentary in both text and speech. It detects key moments like pit stops, overtakes, yellow flags, fastest laps, and builds contextual statements- mimicking the style of live commentators. It can: 1) Generate lap-by-lap summaries 2) Provide position change insights 3) Highlight driver battles 4) Speak commentary aloud using text-to-speech (TTS)
How we built it
We used: 1) Python & Pandas to process and align real-time F1 data (position, laps, race control, weather) 2) OpenF1 API to fetch live/historical race data 3) Natural Language Templates with dynamic insertion logic for generating text commentary 4) pyttsx3 for speech synthesis (TTS) 5) Streamlit for a live interface
Challenges we ran into
1) Handling edge cases like lap overlaps, simultaneous pit stops, and null values in telemetry 2) Generating the right input that ensures that the information provided to the LLM is clean and creates accurate commentary
Accomplishments that we're proud of
1) Created a working pipeline that turns raw data into human-style commentary 2) End-to-End pipeline from web-scraping to front-end streamlit application in under 24 hours
What we learned
1) The power of combining data science with storytelling 2) Challenges of real-time commentary from structured data sources
What's next for BoxBoxBot- AI Generated F1 Commentary
1) Add support for multi-lingual commentary and more natural speech synthesis 2) Build a live dashboard or Twitch overlay for fans and streamers 3) Extend to other motorsports like Formula E or MotoGP
Log in or sign up for Devpost to join the conversation.