I built SnipForge.video completely solo.No co-founder. No team. No VC money.Just Python, Flask, FFmpeg and OpenAI APIs.
Here is exactly how I did it and what I learned.
The Stack
- Python 3 and Flask for backend
- FFmpeg for all video processing
- OpenAI Whisper for AI transcription
- GPT-4 for smart content generation
- Stripe for payments
- SQLite for database
- Vanilla HTML CSS JavaScript for frontend
The entire product runs in a single Python file. Backend, frontend, database, API routes and HTML templates all in one place.
The 28 Tools
Every tool follows the same pattern. Receive
parameters, run FFmpeg or AI API, save output,
update job status.
Tools include trim, compress, convert, crop,
rotate, stabilize, watermark, add captions,
transcribe, extract audio, remove background,
create GIFs, merge videos and more.
The Hardest Part
Auto captions was the most complex feature.
It uses three systems in sequence. FFmpeg extracts audio. Whisper transcribes it with timestamps. Then a custom ASS subtitle file
gets burned back into the video with FFmpeg.
Supporting 20 languages and 4 caption styles while keeping it fast was the real challenge.
Background Job Processing
Every tool runs as a background thread so the browser never waits. The job ID is returned immediately and the browser polls
every 1.5 seconds for progress updates.
This was critical for large file processing.
What I Would Do Differently
Start with fewer tools. I built 28 tools before getting my first user. In hindsight 5 great tools would have been better than
28 average ones at launch.
Try It
Free plan includes 3 videos.Pro is $5 per month for unlimited.
snipforge.video
Use code LAUNCH for 1 month free.
Happy to answer any questions about the technical implementation in the comments.
Top comments (0)