Skip to content

README.md

Latest commit

 

History

History
133 lines (89 loc) · 5.97 KB

README.md

File metadata and controls

133 lines (89 loc) · 5.97 KB

SpectraCleanse AI

Live Pricing Sign up

SpectraCleanse AI strips AI provenance markers and injects SEO metadata to beat algorithmic suppression.

Upload your audio or video file, choose a platform preset, and SpectraCleanse AI will forensically wipe every embedded tag that signals AI-generated origin — then inject clean, platform-optimised metadata powered by Gemini to maximise your reach on YouTube, Spotify, Apple Music, TikTok, and beyond.

Try it free

→ spectracleanse.com

No credit card required. Free accounts include 3 files per month. Upgrade to Creator ($9.99/mo) or Studio ($29.99/mo) for unlimited processing and batch uploads.

How it works

  1. Upload — Quick Cleanse supports MP3 (browser-side). Full Server Cleanse supports MP4 and M4A (server-side). WAV and FLAC are planned but not currently supported.
  2. Analyse — the forensic engine reads every embedded tag and identifies provenance markers.
  3. Cleanse — a nuclear wipe removes all XMP, IPTC, and ID3 fields that could trigger algorithmic detection.
  4. Inject — Gemini generates an SEO-optimised title, description, and tag set tuned to your chosen platform.
  5. Download — receive a clean file with a full forensic report showing exactly what was removed.

Open source

The source code is available at github.com/ChrisAdamsdevelopment/SpectraCleanseAI and is released under the MIT License. You are free to self-host, fork, or contribute.

Local development notes

  • Backend defaults to developer-friendly mode when NODE_ENV is not production.
  • If Stripe env vars are missing locally, /api/create-checkout-session can return a mock checkout redirect (set ENABLE_MOCK_CHECKOUT=true).
  • In production, Stripe variables are still required and the server will fail fast if they are missing.

QA / Deployment Testing

For a step-by-step manual validation flow (local, API smoke, auth, billing, upload, cleanse, Docker, and production readiness), see docs/manual-qa-checklist.md.

  • Browser metadata analysis uses maintained music-metadata with graceful fallback (parseError) when parsing fails, times out, or is skipped for very large files.
  • Quick Cleanse metadata writing remains local/browser-side (MP3 via browser-id3-writer).
  • Full Server Cleanse runs through /api/process; MP3 requests are rejected with HTTP 422. Full Server Cleanse currently supports MP4/M4A only; WAV/FLAC are planned but not currently supported.

Contact

Questions, partnerships, or enterprise enquiries: hello@spectracleanse.com

Native Node deployment runtime

  • Native Render/Node deployments should use Node 20.20.2 (recommended) or another Node 20.x release.
  • If Render defaults your service to a newer Node release, set NODE_VERSION=20.20.2 in the service environment.
  • Node 24 is currently not supported for native installs in validation because better-sqlite3 native compilation failed under Node 24.
  • Docker deployments already pin Node 20 via the repo Dockerfile.

Docker production deployment

This repository includes a multi-stage Dockerfile that builds the frontend and packages dist/ into the final runtime image so server.js can serve the SPA in production.

Build image

docker build -t spectracleanseai:latest .

Run container

docker run --rm -p 3001:3001 \
  -e NODE_ENV=production \
  -e JWT_SECRET=your_jwt_secret \
  -e STRIPE_SECRET_KEY=sk_live_xxx \
  -e STRIPE_WEBHOOK_SECRET=whsec_xxx \
  -e STRIPE_CREATOR_PRICE_ID=price_xxx \
  -e STRIPE_STUDIO_PRICE_ID=price_xxx \
  -e GEMINI_API_KEY=your_gemini_api_key \
  -e FRONTEND_URL=https://your-frontend-domain.example \
  -e DB_PATH=/data/spectra.db \
  -v spectracleanse_data:/data \
  spectracleanseai:latest

Required production environment variables

  • NODE_ENV=production
  • JWT_SECRET
  • STRIPE_SECRET_KEY
  • STRIPE_WEBHOOK_SECRET
  • STRIPE_CREATOR_PRICE_ID
  • STRIPE_STUDIO_PRICE_ID
  • GEMINI_API_KEY
  • FRONTEND_URL
  • DB_PATH
  • REDIS_URL (only if your deployment still uses Redis externally)

Stripe vs local mock checkout

  • Local development may use ENABLE_MOCK_CHECKOUT=true when Stripe variables are not set.
  • Production must use real Stripe configuration; do not rely on mock checkout in production.

Never commit real secrets to source control.

Health & observability

  • GET /api/health (public, no auth) returns { status, uptime, version, time }. Useful for uptime monitors and Render health checks.
  • GET /sitemap.xml and GET /robots.txt are served directly by the backend so they work without the SPA static build (e.g. during dev) and stay in sync with the canonical domain https://spectracleanse.com.

Rate limiting

  • All /api/* routes (except /api/health and the Stripe webhook) are gated by a 60-requests-per-minute-per-IP limiter.
  • /api/login, /api/register, and the /api/auth/* routes get a tighter 10/min limiter on top to make brute-force harder.
  • A throttled request returns 429 with { "error": "Too many requests" }.

Batch processing API

  • POST /api/process-batch (authenticated): processes up to 20 uploaded files sequentially for paid plans (Creator/Studio). Free plan returns 403.
  • GET /api/download/:token (authenticated): one-time secure download for batch outputs.
  • MP3 server cleanse remains unsupported (422 for single process, per-file error in batch); use Quick Cleanse Browser for MP3. Full Server Cleanse currently supports MP4/M4A only; WAV/FLAC are planned but not currently supported.
  • Batch requests enforce a 2GB post-upload soft guard; production deployments should still enforce proxy/body-size/disk limits.