TRY-ON is a next-gen virtual fitting room powered by emotional AI and real-body avatars. It recommends clothing based on real measurements, aesthetic preferences, and personal emotional context.
- Generates a 3D avatar based on your body (manual or scan)
- Matches you with clothing that feels right (not just fits)
- Connects you to brands like LIVER 'IT and others
- Multi-dimensional comparison: fit + emotion + aesthetics
- Open to all brands, with LIVER 'IT as premium showcase
- Data is private, ethical and encrypted
A working HTML/CSS demo is available in the frontend/ folder. Activate GitHub Pages to preview. The landing page now includes a video hero, a horizontal carousel of looks, a 3D avatar preview and an animated contact form to showcase the refreshed UI/UX.
A React version called HeroVideo lives in components/ if you prefer JSX.
The project includes a sample "Casa Pavo Real" fashion line. Shirts, bags,
dresses and hats are defined in products/casa_pavo_real.json and loaded
automatically in the demo.
TRY-ON_APP/
βββ frontend/
β βββ index.html
β βββ css/style.css
βββ backend/ # Ready for Flask or FastAPI
βββ assets/ # Images and logos
βββ docs/ # Technical documentation
βββ README.md
See docs/live_it_codex.md for the full LIVE IT launch codex in Spanish.
- Open
frontend/index.htmlin a browser - Customize avatar flow and clothing logic
- Expand with backend connection later
- Run
npm run buildto generate the static files in thebuild/folder. This is the same command used by Vercel during deployment. - Ensure you have Node.js 18 or newer installed to run the build script locally and match the version used in production.
- Consulta la guΓa de despliegue para mΓ‘s detalles.
- Si usas Netlify y ves el mensaje "Auto publishing locked", sigue estas instrucciones para desbloquearlo.
- Para una explicaciΓ³n rΓ‘pida en espaΓ±ol sobre cΓ³mo compilar y ejecutar la demo local, revisa docs/demo_local.md.
Run npm run build to generate the static build/ directory used by Vercel. This copies the main HTML, CSS and assets so the site can be deployed anywhere.
Run node scripts/generate_vercel_env.js to create a .env file with your Vercel organization and project IDs. Set VERCEL_TOKEN beforehand so the script can query the API.
python scripts/codex_deploy_automation.py generates the .env file and runs make all.
Ensure the environment variables from .env.example are exported before executing it. The script connects Make, Codex (GPT-4) and Vercel for a full automated deploy.
Run ./export.sh to bundle the website and demo app into TryonApp_export.zip. You
can then upload the ZIP archive to Google Drive or any file-sharing service to
share the LIVE 'IT and TRY-ON demo.
See docs/shopify_domain_setup.md for instructions
on connecting tryonyou.app to Shopify and uploading the storefront files.
To remove the storefront password automatically, follow
docs/unlock_shopify_password.md.
Access the project via the following subdomains:
https://tryonyou.appβ Wix landing pagehttps://app.tryonyou.appβ The active TryOnMe apphttps://api.tryonyou.appβ Node.js API connected to Firebase and Shopifyhttps://gpt.tryonyou.appβ Custom AI agents
MIT License β Feel free to fork, remix, and try it on.
The project includes modular AI agents located in the agents/ folder. Import them via require('./agents') to access:
- DeployScriptMaster: deploys to Vercel.
- TestRunnerGPT: sets up and runs tests.
- RefactorAgentJS: simplifies JavaScript functions.
- DocuGen: generates documentation snippets.
- TaskResolver: provides quick task actions.
- compararTalla: compara medidas y devuelve la talla Γ³ptima.
- recomendarPrenda: sugiere la prenda ideal.
- reaccionPau: genera reacciones de Pau.
- validarDonacion: comprueba si una prenda puede donarse.
- generarAvatar: crea un avatar 3D personalizado.
- brandGuardian: asegura la coherencia de la marca.
Para una introducciΓ³n rΓ‘pida a estos agentes consulta docs/agentes_basicos.md.
Set the TRYON_PLAN environment variable to control which modules are active.
Available plans are basic, pro and ultra. Premium scripts like fit_ai.py
will exit if the required module is not enabled for your plan.
Run python scripts/trending_products.py to fetch trending searches in Google Trends for the Apparel category and generate prompt ideas.
Run python scripts/generate_seo_meta.py to create products/seo_meta.json with titles and descriptions derived from your product list. The script now reads from products/tryon-products.json by default.
Run python scripts/generate_image_meta.py to create products/image_meta.json with alt text and social image tags for each product. It also defaults to products/tryon-products.json.
Run python scripts/dropshipping_scraper.py to gather factory offers from data/dropshipping_sites.json into dropshipping_report.csv.
Use python scripts/sync_to_shopify.py to upload products defined in products/tryon-products.json (or products/casa_pavo_real.json as a fallback) directly to your Shopify store. The script expects SHOPIFY_STORE_DOMAIN and SHOPIFY_ACCESS_TOKEN environment variables for authentication. A Node.js variant is available via node scripts/upload_to_shopify.js using the same variables.
Run python scripts/auto_update.py --once to fetch the demo dataset and optionally sync products. Add the --interval N flag to run it every N minutes or keep it running continuously. Set ENABLE_SHOPIFY_SYNC=1 in the environment to push new products after each fetch.
Run python scripts/send_push.py USER_ID "Your message" to send a notification via OneSignal. Supply ONESIGNAL_API_KEY and ONESIGNAL_APP_ID as options or environment variables.
Automation flows for the Automa extension can be found in docs/automa_flows.md.
Scenarios for the Make platform are listed in docs/make_scenarios.md.
For instructions on automating Vercel deploys with Make and OpenAI, see docs/deploy_vercel_make.md.
For a combined workflow using both tools, check docs/automa_make_flow.md.
Run python scripts/site_health_check.py to verify that the main site is reachable.
If the check fails and email credentials are configured via SMTP_HOST,
SMTP_USER, SMTP_PASS and NOTIFY_EMAIL, an alert will be sent automatically.
El archivo docs/flujo_automatizado_semanal.md detalla una agenda semanal de agentes para coordinar productos, pedidos y contenido.
Use python scripts/instagram_story.py <imagen_url> <nombre> <precio> --token YOUR_TOKEN to post product stories to your Instagram account. The token can also come from the INSTAGRAM_ACCESS_TOKEN environment variable. See docs/instagram_story_autopost.md for details.
Run python codex.super_executor.py to launch a minimal bot that triggers multiple virtual agents in parallel (PMV, HR Supervisor, Tester UX Web and others). Each agent prints its progress so you can verify the coordination.
The folder scripts/tryonme_full contains a minimal implementation that combines a Node.js helper to generate TryOnMe profiles with GPT-4, a Flask API to create or search profiles and an embeddable HTML interface. Install dependencies with npm install and pip install -r requirements.txt. Launch the Flask app via python scripts/tryonme_full/flask_app.py and open scripts/tryonme_full/embed.html in a browser.
Consulta docs/financial_infrastructure.md para detalles sobre la capa de pagos y recompensas basada en Revolut Ultra.
See the deployment guide for more details. For a quick checklist of what to verify before going live, read docs/deployment_review.md. For a condensed reference that links all deployment docs, see docs/deployment_overview.md.
If you use Netlify and see "Auto publishing locked", follow these instructions to unlock it.
Run npm run build to generate the static build/ directory used by Vercel.
This copies the main HTML, CSS and assets so the site can be deployed anywhere.