NivoSearch

01Overview

NivoSearch is an independent web search product I'm building from the ground up. It covers the full vertical stack, crawling, indexing, relevance ranking and a results interface, rather than wrapping someone else's API.

02The Challenge

Search is hard because quality and cost live on opposite sides of the same system. The challenge is keeping the index current and the ranking honest without infrastructure costs spiralling out of control before the product has revenue.

03What I Built

• A crawling and ingestion pipeline that fetches, parses and normalises content into a structured index.
• A ranking layer that blends relevance signals, freshness and simple behavioural feedback.
• A results interface with the kind of restraint a search product needs, nothing decorative.
• Monitoring and reporting so the index health is visible at all times.

04Technical Architecture

• →Crawlers enqueue work into a queue and write raw content to object storage.
• →A processing stage parses and extracts content, then indexes it in batches.
• →The query API sits in front of the index and applies the ranking model.
• →The frontend is a thin client over that API, server-rendered where it matters.

05Technologies

06My Role

Solo builder, product, architecture, implementation and operations.