01 / The problem
What the system needs to solve
Turning voice notes into short videos meant transcribing, writing subtitles, choosing background footage and music, and rendering by hand for every clip.
02 / The approach
How the pieces fit together
- Receive voice or audio input from users through a Telegram bot.
- Transcribe speech with the ElevenLabs API and generate SRT subtitles.
- Select background video and music dynamically from Google Drive.
- Render over SSH with FFmpeg (subtitle overlay and audio mixing), then store and deliver the result.
03 / Key features
What it does
- Complete pipeline: input → processing → enhancement → output → delivery
- Fallback delivery links when files exceed size limits
- Automated file handling and storage in Google Drive
04 / The tradeoffs
Engineering is a set of choices
Rendering with FFmpeg on a server over SSH keeps n8n lightweight and the orchestration easy to inspect, but splits state across two systems, so file handoffs and failures need explicit handling.
05 / Validation considerations
What to test before relying on it
Test oversize files, failed transcriptions and missing media, and confirm every request ends in a delivered video or a clear fallback link.