Fera-Search – Privacy-First AI Meta Search Engine

Name: Fera-Search – Privacy-First AI Meta Search Engine
Availability: InStock
Author: himanshu_fera

by himanshu_fera·Feb 19, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidShip ItDark Horse

The Take

The UX choice to return search results immediately and run Sarvam summaries in the background is the project's clearest win — perceived latency stays low while AI adds value. Under the hood it's mostly orchestration: a SearX-backed meta-search with a FastAPI proxy on Hugging Face and a GitHub Pages front end, which makes it easy to audit and self-host but not revolutionary. Worth trying if you care about no-login privacy, though you should scrutinize where summary generation runs and whether the proxy leaks queries.

Post Description

Hi HN,

I built Fera-Search as a privacy-focused meta search engine with AI summaries.

The goal was to combine: - Multiple upstream search engines - Minimal tracking - Fast response times - AI summaries that run separately so results stay instant

Architecture: - Frontend on GitHub Pages - FastAPI proxy on Hugging Face - SearX upstream engines - Sarvam for structured AI summarization

Search results return immediately, and AI runs in the background so the UI never blocks.

This is an early version and I’d really appreciate feedback on: - UX improvements - Performance - Privacy model - Feature ideas

Thanks for taking a look.