InsideFi — Call Intelligence Platform

$ cat ./problem

The problem

Pay-per-call marketers run on phone calls the way e-commerce runs on clicks — but their data lives locked inside dialers. Ring-level detail sat in Ringba, caller history didn't exist anywhere, call outcomes were judged by gut, and billing partners meant spreadsheets. The team needed the whole revenue loop — traffic in, insight out, invoice sent — in one platform.

$ git log --built

What we built

• Ring-level ingestion pipeline syncing continuously from Ringba — every ring, normalised and queryable
• Caller-analysis profiles built across campaigns: history, quality signals and repeat behaviour per caller
• AI disposition reports that read call outcomes and summarise what actually happened on the line
• Coverage analysis and routing insight feeding the visual journey builder
• Billing & invoicing on top of the same data — the numbers that get invoiced are the numbers that were measured

$ cat ./architecture

How it's put together

Edge-first by design: Hono services on Cloudflare Workers at the front, ClickHouse doing the heavy aggregation over event streams, Neon Postgres for transactional records, Redis for the hot path. The result is analytics that stay fast as ring volume grows.