Seven hundred pitch decks,
one search engine that
actually
finds them.
Funding Fyre is a resource hub for founders — 700+ globally pitched decks, investment memos, and pitch reels, indexed by industry, funding stage, and round. The platform's value depends on a single question: can a founder find the exact reference they need in under five seconds?
We rebuilt the platform from infrastructure up — a scalable cloud delivery layer, a separate document store for resilience, ML‑assisted search, magic‑link onboarding, and a UI that gets out of the way of the material itself.
The Challenge
Volume that has to feel small.
A 700‑deck library is only useful if the right deck surfaces in seconds. Pile that together with investment memos, pitch reels, and thousands of metadata fields, and the design problem becomes a search problem dressed up as a UI problem.
The platform also has to stay up under spiky traffic, protect a curated corpus, and give first‑time visitors a sense of scale without making them dig through a wall of links.
Founders read pitch decks for inspiration. They don't have time to fight an interface to do it.
What Webloon Did
Infrastructure, search, and an interface that disappears.
We treated this as four interlocking sub‑projects: cloud delivery, document store, search and ranking, and the UI on top. Each one solved a different bottleneck — and together they let the catalogue scale without the experience getting heavier.
Architecture
Cloud delivery
- AWS‑based delivery with CDN front
- Separate MongoDB document store for resilience
- Capacity headroom for catalogue growth
Search
Ranking & filters
- Filters by industry, stage, funding round, and name
- Machine‑learning ranking that improves with use
- Indexing pipeline for new resources
Design
Interface
- Clean, minimal grid for browsing
- Deck reader with focused reading mode
- Interactive guides for first‑time visitors
Security
Onboarding & auth
- Magic‑link login as primary auth
- Database‑layer isolation for risk mitigation
- Edge protections and rate limiting
What This Proves
A library that performs like a product.
Post‑launch metrics confirmed the architecture choices were right: average session time and search efficiency both improved, near‑zero downtime under spiky load, and the catalogue continues to grow without re‑platforming.
Quantitative results available on request under engagement confidentiality.