Startup aggregator with self-service profile pages and AI-assisted quality control
The aggregator runs the main site and moderation, while each startup gets its own page and secure access to edit only its own profile. An AI layer raises catalog quality, speeds up moderation, and lifts CTA conversion across promo packages.
About the Project
A startup aggregator that solves two jobs in one product. For visitors, it's a searchable directory with filters: find a relevant startup, open the profile, see everything that matters in one place. For startups, it's a profile inside a shared template — description, tags, media, team, links, supporting materials, CTAs. The core idea was to take catalog maintenance off the aggregator's team without sacrificing quality: startups update their own content, but only through clear rules, publishing statuses, and moderation.
About Customer
Jules is an entrepreneur from Lyon. He's building a startup aggregator that monetizes attention and access — promoted listings in the directory, promo packages, partnerships, investor and client nurturing, and a layer of additional services around the ecosystem. For him, a steady brand and a clean catalog are the product. In a directory like this, trust is the asset that gets eroded the moment pages start varying in quality, data goes stale, or the tag tree turns chaotic. Jules came to us through a recommendation from the business community, looking for a long-term technical partner to grow the product with him.
Brand website plus directory
The project had to cover both the public-facing aggregator experience — showcase pages, sections, the directory itself — and the startup content layer, with profiles, dictionaries, and media — all reading as one coherent product.
Self-service editing
Startups needed to update their own content without any access to other startups' pages, dictionaries, or the catalog structure. A single permissions mistake is a direct trust risk.
Catalog scalability
Once the directory grows into the dozens and hundreds, hand-building pages destroys the aggregator's economics. The template and data model had to support growth without one-off custom handling.
SEO as a product feature
Every profile is its own organic entry point. The platform needed stable URLs, controllable metadata, predictable indexing, fast page speed, and quality controls that keep pages from degrading over time.
Two audiences
Visitors want to find and compare quickly. Startup representatives want to update a profile fast and clear moderation. Both flows had to stay simple.
Catalog quality
Promo packages, partnerships, and Featured placements sell only when the pages look clean and consistent. The product needed rules, moderation, and mechanics that actively help startups fill out profiles correctly.
AI to strengthen the process
AI had to support quality — suggest tags, identify gaps, propose stronger copy — without inventing facts or quietly publishing on its own.
Proposed Solution
We built the platform as a headless system: Directus as the back office and source of truth for the directory, dictionaries, moderation, and access, with Next.js as the fast public site and SEO-friendly rendering layer. Responsibility was split cleanly — the aggregator owns structure and quality, startups have self-service editing inside a managed template.
Public website on Next.js and startup profiles
We built the aggregator's central pages, the directory, the profile pages, and the base navigation. On top, the structure was prepared for filters, search, and the fast user journey: find → open → click the CTA.
Directus as a managed back office with multi-tenant access
We modeled the collections for startups, categories, tags, media, team members, materials, promo statuses, and the moderation entities. Permissions were configured so a startup representative sees and edits only their own record, while the aggregator team owns dictionaries, rules, statuses, and moderation.
Managed startup-page template
We defined a set of blocks and fields that let profiles differ in content while staying consistent in quality. The structure covers short description, value proposition, industry and stage, tags, team, links, materials, and CTAs — freedom inside a framework that protects the brand.
Moderation flow
We introduced statuses (Draft → In Review → Published → Hidden), along with checks and control fields that block empty or invalid profiles from going live.
AI layer for quality and scale
AI was built as an assistant, never a publisher. It proposes a sharper short description and value proposition, recommends tags and categories, generates a gap-checklist before moderation, and drafts short promo theses for Featured placements and partner collections. Every AI suggestion has to be reviewed and approved by the startup or the moderator — nothing publishes itself.
SEO foundation for catalog growth
Stable URLs, metadata, indexing rules, SSR / SSG where it counts, and content-quality controls so profile pages work as durable organic entry points years from now.
We started by defining the rules and tensions inside the aggregator model — startups want visibility and leads, the aggregator monetizes promo and partnerships, visitors won't tolerate a chaotic catalog. From there we defined the minimum profile standard every startup had to hit and the field set critical for search, comparison, and SEO. In parallel, we defined which parts of the catalog had to stay strictly controlled (dictionaries, statuses, moderation) and where freedom was fine (descriptions, media, supporting materials).
+2 Resource
Results
The metrics below land directly on the aggregator's monetization levers — promo, partnerships, visibility — and on the operating economics underneath them.
10–15 minutes
to onboard a new startup
Under 24 hours
target moderation SLA for profiles and profile updates through the Directus queue
60–75% of profiles
clear moderation on the first attempt after the AI quality checklist
70–85% fewer
manual profile corrections by the aggregator team
120 new SEO pages
appear on the platform every month, with no developer involvement
+20–35% CTR
on profile-page CTAs after profile-structure standardization and AI-assisted CTA suggestions
Featured collections
prepared 2–4× faster
Zero risk
of editing someone else's profile, or having yours edited by another startup
Get in Touch
Or send us a message and we'll get back to you within 15 minutes during business hours