AI Campaign & Product Page Generator for eCommerce
A configurable workflow that turns a brief into a live page and deploys it directly to your CMS or store platformIt covers the product launches, promotions, and seasonal campaign pages a retail or ecommerce business publishes continuously, so routine page production does not require a designer for every request
From a brief to a live campaign page, automatically.
The situation
Retail and ecommerce businesses publish product launches, promotions, and seasonal pages continuously. Built by hand, each one took hours; the designer and copywriter became the bottleneck for every campaign, brand consistency drifted from page to page, and urgent requests queued for a day or two.
How we built it
A configurable n8n workflow that takes a brief from Telegram, a web form, or Slack, writes copy consistent with the defined brand voice and content rules, generates the visuals, assembles the page, and deploys it to the team's platform. The workflow is built around the systems the team already uses: models, image providers, and deployment targets can each be replaced independently.
What it does
Multi-mode routing
A switch picks the mode for each request: full generation, code only, or images only.
Brand-consistent copy
The model writes from the team's profile (brand voice, tone, legal constraints) and can be replaced with GPT-4o, Gemini, or a self-hosted model.
Auto image generation
Hero image, logo, and favicon are generated automatically; the image provider is swappable for Midjourney, DALL-E, or Stable Diffusion.
Pluggable deployment
Deploy to Keitaro, WordPress, Webflow, Framer, S3 or Cloudflare, a custom CMS, or any REST API, configured for the team's stack.
HTML assembly
A code step assembles the final HTML and CSS and inlines assets into a single self-contained file.
Live notifications
Real-time deploy status to Telegram, Slack, or email, with the link, build time, and mode.
Technical scale
- 15+
- Workflow nodes
- 3
- Generation modes
- 6+
- Deploy targets
- 4+
- Swappable model options
What it runs on
Orchestration
n8n with Webhook, Switch, and Code nodes, self-hosted or on n8n Cloud
AI models (swappable)
Claude, GPT-4o, Gemini, or a self-hosted model
Image generation (swappable)
Flux, Midjourney, DALL-E, or Stable Diffusion
Deployment (pluggable)
Keitaro, WordPress, Webflow, Framer, S3 or Cloudflare, REST API
Storage & monitoring
Supabase, Postgres, or MySQL; status to Telegram, Slack, email, and n8n logs
A deeper breakdown of the architecture and modules is available on request. There is too much detail for a public page.
What changed
- A page request in chat becomes a deployed page without a ticket, and routine production does not require a designer for every request.
- One brand profile per team keeps generated pages consistent with the defined brand voice and content rules.
- Additional A/B variants can be produced with substantially less manual production effort.
- The team can publish many more pages per day at the same headcount.
- Models, image providers, and deployment targets can be replaced independently, so the workflow fits the systems already in use.