AI-Powered Content Marketing Platform with Cloudflare & Real-Time Collaboration
In today's crowded digital landscape, marketing teams need to create high-quality, on-brand content and publish it across multiple channels — often within minutes, not days. Businesses that rely on generic AI tools struggle with inconsistent brand voice, disconnected workflows, and manual handoffs between writing, review, and publishing. Our client, a growing B2B SaaS platform for marketing teams, was facing exactly this challenge: content creation was slow, brand guidelines were hard to enforce at scale, and publishing to social and CMS platforms required too many steps and too much context-switching.
We partnered with the team to design and build Neonify — an AI-powered, collaboration-first marketing platform deployed on Cloudflare's edge. The solution unifies brand voice management, AI content generation, campaign planning, knowledge ingestion, and multi-platform publishing in a single workspace. By combining Cloudflare Workers AI, durable workflows, and real-time collaborative editing, the platform now helps teams go from idea to published post in a fraction of the time. The project eliminated fragmented tooling, improved team collaboration, and created a flexible architecture that adapts as integrations and AI capabilities grow.
Project Info
Client
Neonify (B2B SaaS — AI marketing & content platform)
Industry
Marketing technology / MarTech
Date
June 2026
Platform
neonify.aiStack
Next.js 15, Cloudflare Workers, Workers AI, Liveblocks, Clerk
Background
Neonify serves marketing teams and agencies who need to produce consistent, high-quality content aligned to their brand voice and publish it across social media and CMS platforms. The platform targets organizations that want AI assistance without sacrificing brand control or team collaboration.
Before the current architecture and product refinements, the application faced several structural limitations: navigation lived behind a fake single-page app where every section shared the URL `/`, making pages impossible to bookmark or share; post-sign-in flows suffered from session race conditions; social OAuth and integrations were scattered across multiple screens; AI-heavy workloads were split across multiple Cloudflare Workers; and application state leaned heavily on Liveblocks as a primary data store rather than a dedicated database.
The goal was to build a faster, smarter system that could capture brand knowledge, generate personalized on-brand content, enable real-time team editing, and deliver posts directly to connected platforms — all on serverless edge infrastructure that scales with demand.
The Challenges
Fragmented User Experience – Users navigated through a client-side SPA layered on Next.js where URLs never changed, breaking browser history, shareable links, and per-route middleware protection.
Auth & Session Friction – Soft client-side redirects after sign-in caused race conditions with Clerk session propagation, leaving users in broken loading states instead of the dashboard.
Slow Path to First Post – Onboarding required too many steps before users could connect a social account and publish. Without an early integration, the core product promise was unreachable.
Inconsistent Brand Voice – Generic AI outputs didn't reflect each organization's tone, style, and expertise. Teams needed a system that learns from their documents, websites, and guidelines.
Disconnected Publishing – Content lived in one place; Facebook, Instagram, LinkedIn, and WordPress connections lived elsewhere. Users constantly left their current screen to finish the next step.
Scalability & Architecture Debt – Long-running AI jobs, document processing, and cron tasks were spread across multiple workers with HTTP fallbacks, duplicate R2 buckets, and post-build injection hacks.
Knowledge Silos – Marketing teams couldn't easily ingest URLs, files, images, or videos into a searchable knowledge base to inform AI-generated content.
Solution
We designed and implemented an AI-powered, edge-deployed marketing platform using modern serverless patterns and a collaboration-first architecture — unifying brand voice, AI generation, real-time editing, knowledge ingestion, and multi-channel publishing in one workspace.
Brand Voice & AI Content Generation – Automated brand profiling from documents and URLs, AI-enhanced writing aligned to each organization's voice, platform-specific previews, and EEAT analysis for SEO-quality output.
Real-Time Collaboration – Liveblocks-powered Lexical editor with presence, cursors, and live updates. Multi-workspace support per Clerk organization for agencies managing multiple clients.
Campaign & Publishing Workflow – Unified campaign calendar, Quick Post flow for faster time-to-first-publish, and OAuth-based publishing to WordPress, Facebook, Instagram, and LinkedIn.
Knowledge Base – Multi-source ingestion from URLs, websites, PDFs, DOCX, images, and video with semantic search powered by Cloudflare AI Search and durable Workflows for long-running jobs.
Unified Integrations Hub – Consolidated `/social-media`, `/cms`, and `/email` routes into a single `/integrations` destination with platform-specific OAuth and Facebook page sync.
Edge Platform & Infrastructure – Next.js 15 App Router with real URLs, OpenNext deployment on Cloudflare Workers, Clerk auth and billing, server actions with next-safe-action + Zod, and R2 for asset storage.
Auth & Routing Fixes – Hard navigation post-sign-in eliminates Clerk session race conditions. SessionReadyGate waits for session, workspace, and plan limits. Canonical routes in `lib/routes.ts` serve as a single source of truth.