Experience layer
Cloudflare Pages
Host the public product surface: homepage, Playgrounds, Workflows, Stack pages, Games, and SEO landing pages.
Deploy the Next.js front end and mostly static MVP pages with minimal operational work.
Cloudflare-first AI infrastructure
A low-ops stack for building, deploying, and operating AI-native products with Pages, Workers, Workers AI, D1, KV, R2, Cron Triggers, and analytics.
Architecture map
Experience layer
Host the public product surface: homepage, Playgrounds, Workflows, Stack pages, Games, and SEO landing pages.
Deploy the Next.js front end and mostly static MVP pages with minimal operational work.
API layer
Run lightweight APIs for generation requests, submit forms, sponsor leads, public result pages, and workflow pack requests.
Keep product logic close to the edge while avoiding a traditional server.
AI runtime
Generate product plans, workflow drafts, SEO briefs, benchmark summaries, and story branches after deterministic MVP templates are validated.
Power AI Product Factory outputs and future Conscious Worlds episode generation.
Structured data
Store structured records such as product submissions, sponsor leads, saved results, workflow requests, and benchmark runs.
Use the existing `richbayai-db` as the low-maintenance SQL base for the MVP data layer.
Fast state and cache
Cache public generated outputs, feature flags, page settings, and repeated low-risk reads.
Speed up public result pages and generated content while keeping D1 for canonical records.
Assets and downloads
Store generated screenshots, story images, workflow pack files, benchmark exports, and other binary assets.
Host future downloadable packs and generated media without adding a separate storage provider.
Automation
Run scheduled AI-assisted update jobs for content refresh, benchmark candidates, tool checks, and weekly monitoring.
Support the AI-generated, human-reviewed publishing loop.
Measurement
Track generation runs, copy/share/remix clicks, workflow CTA clicks, submit leads, and sponsor intent.
Decide which Playgrounds, Workflows, and Stack pages deserve more investment.
Use cases
Generate plans, cache result URLs, save high-intent runs, and route users into Workflows.
Generate future story branches, save endings, and serve lightweight assets.
Store pack requests, delivery metadata, and prompt/template versions.
Track affiliate/sponsor fit and keep stack recommendations structured.
Store scenario inputs, model outputs, scoring notes, and update timestamps.
Structured data
playground_runsGenerated result metadata, input summary, output snapshot, share slug, locale, and CTA clicks.
submissionsAI product submissions from `/submit`, review status, category, URL, and contact email.
sponsor_leadsSponsor intent, placement type, budget range, target page, and contact status.
workflow_requestsWorkflow pack interest, requested pack, source page, and lead email when provided.
benchmark_runsScenario, model/tool tested, criteria, result summary, score, and reviewer notes.
generated_pagesAI-assisted page drafts, review state, target keyword, published URL, and refresh date.
How to ship safely
Keep pages static, validate positioning, build the Playground -> Workflow -> Stack chain.
Persist submit leads, sponsor leads, and selected Playground runs in D1.
Replace deterministic result templates with Workers AI where output quality is proven.
Add cached public result pages, generated screenshots, and downloadable workflow packs.
Use Cron Triggers to draft updates, then require human review before publishing.