🚄 Subway Surfers Text-to-Video Converter

Skip the fairy-tale, just give me the code already --> Click Here

Inspired by the absolute legend danielbonkowsky/subwaysurfers-text, this project throws some sauce on the original concept. If this container saves you time (or your sanity), go drop a star on his project already, damn it.

🚀 What the Hell Is This?

You ever try to focus on an article and your brain taps out after the second paragraph? Same. This little monster extracts text from a webpage and slaps it onto hyperactive Subway Surfers gameplay footage—turning dry, boring content into something your scrolling-addicted brain can actually sit through. Perfect for TikTok, ADHD, or just making studying suck less. You get:
🧠 smart content + 🎮 mindless gameplay = ✅ actual retention (probably)

📸 Screenshot

🎯 Oh Hell Yeah, Features

• Converts your average web article into a Subway Surfers-screaming-TikTok-style video
• Rips text straight from URLs
• Slaps it on top of fast-ass gameplay
• Bonus points: speech-to-text powered by Vosk API
• Built with Flask + OpenCV + goose3 + Python + zero chill

🐳 Running This Beast with Docker

Step 1: Pull the image, because we're not building this crap from scratch:

docker pull tebwritescode/subwaysurfers-text20:latest

Step 2: Get the Vosk brain model and unzip it:

wget https://alphacephei.com/vosk/models/vosk-model-en-us-0.22.zip
unzip vosk-model-en-us-0.22.zip -d ./
rm vosk-model-en-us-0.22.zip

Step 3: Drop a gameplay video into ./sources — any .mp4 will do. More videos = more fun. Only one? It'll reuse that sucker on every run.

Step 4: Let's ride:

docker run -d --name subwaysurfers-text \
  -v ./subwaysurfers/sources:/app/sources \
  -v ./subwaysurfers/vosk-model-en-us-0.22:/app/static/vosk-model-en-us-0.22 \
  -p 5000:5000 \
  tebwritescode/subwaysurfers-text20:latest

Step 5: Fire up your browser and hit:

http://localhost:5000

Get ready to weep tears of productivity as your article runs across a CGI train track.

🔧 Environment Variables For Control Freaks

Want to make this thing dance to your tune? Here are the variables you can tweak:

# Path to the Vosk TTS model
MODEL_PATH=static/vosk-model-en-us-0.22

# Where your Subway Surfers gameplay videos live
SOURCE_VIDEO_DIR=sources

# Pick a random source video each time? 
# Set to "false" if you want to use a specific file for SOURCE_VIDEO_DIR(sources/myfile.mp4)
RANDOM_SOURCE=true

Add them to your docker run command like a boss:

docker run -d --name subwaysurfers-text \
  -e MODEL_PATH=static/custom-model \
  -e RANDOM_SOURCE=false \
  -e SOURCE_VIDEO_DIR=sources/myfile.mp4 \
  -v ./subwaysurfers/sources:/app/sources \
  -v ./subwaysurfers/vosk-model-en-us-0.22:/app/static/vosk-model-en-us-0.22 \
  -p 5000:5000 \
  tebwritescode/subwaysurfers-text20:latest

Or in your docker-compose like a champion:

services:
  subwaysurfers-text:
    image: tebwritescode/subwaysurfers-text20:latest
    environment:
      - MODEL_PATH=static/vosk-model-en-us-0.22
      - SOURCE_VIDEO_DIR=sources
      - RANDOM_SOURCE=true
    volumes:
      - <HOSTPATH>/sources:/app/sources
      - <HOSTPATH>/vosk-model-en-us-0.22:/app/static/vosk-model-en-us-0.22
    ports:
      - "5000:5000"

🛠 Using Docker Compose Like a True King 👑

Replace <HOSTPATH> with the full path to where your god-tier video and Vosk model live.

services:
  subwaysurfers-text:
    image: tebwritescode/subwaysurfers-text20:latest
    volumes:
      - <HOSTPATH>/sources:/app/sources
      - <HOSTPATH>/vosk-model-en-us-0.22:/app/static/vosk-model-en-us-0.22
    ports:
      - "5000:5000"

Start it up:

docker-compose up -d

Fun fact: if you only mount one .mp4, the app will always use that. Mount a folder of them? It picks a random vid each time. 🎲

🐍 The Dockerfile Straight-Up Breakdown

• Uses a slimmed-down official Python image
• Installs all the nasty dependencies (yes, even ffmpeg and build tools)
• Brings in Homebrew in case you're feeling fancy
• Copies the app, installs Python and brew packages
• Runs Flask like it means it
• Your container, your rules

🔮 Shit I Still Wanna Add

• Drop-down to select different AI voices (because not every AI wants to sound like Siri on Ambien)
• ~NVIDIA GPU acceleration (because some of you have beefy rigs)~ Nixed this shit, GPU transcoding is slow AF
• ~Fix that annoying subtitle-stuck bug nobody asked for~ I think I squashed this fucker

🧠 Want to Run It Without Docker?

You rebel. Here's the repo for bare-metal installs:
👉 https://github.com/tebwritescode/subwaysurfers-text-multi

🏆 Big-Ass Credit Section

Massive props to danielbonkowsky for the core idea. This is just a remixed, Dockerized, slightly louder sibling. Go say thanks.

⚠️ Final Note from His Royal Freshness™

Yeah, the container works. It does the job. But it's packing a little extra weight and could use a clean-up. If trimming containers gives you a tech boner, PRs are wide open. Enjoy learning like a hyper-caffeinated teenager on TikTok 🚀