In brief
NeuralKite maps private companies in fragmented UK sectors, scores targets against an acquisition thesis and produces decision-ready dossiers. Its workflows run from ownership discovery through screening, thesis-fit scoring and shortlisting to export; the sectors named on the live site are UK MSPs, accountancy firms and IT services.
Under the surface it is a Next.js App Router application in TypeScript, with Prisma on PostgreSQL, NextAuth, Stripe billing, LLM agents on the Anthropic SDK for company research and dossiers, Playwright and Cheerio scrapers, and server-sent events for live discovery. It sits on a self-hosted PostgreSQL warehouse that includes a 2.8-million-company Crunchbase dataset I loaded and analysed.

Status
Status note
Live product, deployed and publicly accessible; customer numbers, revenue and launch stage are not published, and none are claimed here.
The full case study, covering the problem, my contribution, the technical decisions, the evidence, the limitations and what I learned, follows.
Links
- neuralkite.com (external link)Public landing page with sign-in and a demo booking.