Skip to content

hasankaantan/SceneFXai

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

457 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SceneFX AI

AI-powered sound design for video creators.
Upload an SRT subtitle file → Claude AI analyzes every scene → ElevenLabs generates royalty-free SFX → MusicGen creates background music → auto-mixed at −14 LUFS YouTube standard.

🔗 Live: scenefxai.app


What It Does

  1. SRT Analysis — Claude AI reads each subtitle line, groups them into scenes, and identifies emotion, tempo, and audio needs per scene.
  2. SFX Generation — ElevenLabs generates scene-specific sound effects with automatic timestamp placement.
  3. Music Generation — Replicate MusicGen creates mood-matched background music for each scene.
  4. Auto Mix — FFmpeg mixes original audio + SFX + music at −14 LUFS, normalized for YouTube.
  5. Download — One-click MP3 download.

Result: 4–8 hours of manual sound design → ~3 minutes, fully automated.


Tech Stack

Layer Technology
Frontend Next.js 16 · React 19 · TypeScript · Tailwind CSS
Backend Next.js API Routes · BullMQ Workers
Database Supabase (PostgreSQL)
Queue / Cache Upstash Redis
Storage Cloudflare R2
AI — Analysis Anthropic Claude (claude-sonnet-4-6)
AI — SFX ElevenLabs Sound Effects API
AI — Music Replicate / MusicGen (stereo-large)
Audio Processing FFmpeg (via Railway workers)
Auth Supabase Auth (GoTrue)
Payments Stripe
Email Resend
Monitoring Sentry
Hosting Vercel (Next.js) · Railway (workers)
CI/CD GitHub Actions · Gitleaks · Dependabot

Architecture

Browser (Vercel)
  └── Next.js API Routes
        ├── srt-queue     → srt-worker (Railway)
        │                     └── Claude SRT analysis + ffmpeg audio extract
        ├── audio-queue   → audio-generation-worker (Railway)
        │                     └── ElevenLabs SFX + Replicate MusicGen
        └── mix-queue     → mix-worker (Railway)
                              └── FFmpeg mix → Cloudflare R2 upload

All queues: BullMQ over Upstash Redis
All audio files: stored in Cloudflare R2

Business Model

Credit-based SaaS (pay-as-you-go):

Package Price Credits
Starter $9 200 cr
Pro $19 500 cr
Studio $35 1,000 cr
  • SRT Analysis: 5 credits
  • SFX per scene: 5 credits
  • Music per scene: 5 credits
  • Gross margin: ~85–88%
  • Monthly fixed infra cost: ~$31

Security

  • Security score: 91/100
  • Auth & Authorization: 20/20 (Supabase Auth + RLS)
  • Rate Limiting: 10/10 (Upstash)
  • Input Validation: 18/20
  • Monitoring: 10/10 (Sentry)
  • Secret Management: 15/15 (env vars, Gitleaks)

Running Locally

# Install dependencies
npm ci

# Copy environment variables
cp .env.example .env.local
# Fill in: ANTHROPIC_API_KEY, ELEVENLABS_API_KEY, REPLICATE_API_KEY,
#          UPSTASH_REDIS_URL, SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY,
#          CLOUDFLARE_R2_*, STRIPE_*, RESEND_API_KEY

# Start Next.js
npm run dev

# Start workers (separate terminals)
npm run worker:srt
npm run worker:audio
npm run worker:mix

Contact

Fatih Kunakfatih.kunak@gmail.com
scenefxai.app

About

Resources

License

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors