Nicheloom

Market intelligence for builders — see what's gaining traction before it's crowded.

ParqDB

Vector search in the browser from Parquet over HTTP"

Details

External ID
49382022
Source
HN
Company
—
Product
ParqDB
Website domain
parqdb.io
Launched
Aug. 21, 2026
Cohort
—
Upvotes
26
Upvotes percentile
0.7916666666666666
Tags
—
Fetched at
Sept. 10, 2026, 5:32 a.m.
Updated at
Sept. 10, 2026, 5:32 a.m.

Enrichment

Theme
web development and browser utilities
Vertical
Horizontal
Function
Search & retrieval
Audience
Developer
AI stance
Not AI
Project type
Commercial product
Normalized one-liner
vector search in browser from parquet over http
Manually corrected
False

Could you build this?

Partial Querying Apache Parquet files directly over HTTP range requests and running in-browser vector search requires low-level byte parsing and SIMD-optimized vector math in WebAssembly.

What it would actually take: The client uses WebAssembly (compiled from Rust or C++) to issue HTTP range requests fetching Apache Parquet headers and vector chunk payloads, avoiding full file downloads. The core difficulty lies in writing an efficient vector similarity index (e.g., HNSW or IVF) that works within browser memory limits using WebAssembly SIMD without server-side compute. Building this requires deep knowledge of columnar storage formats, binary byte manipulation, and vector indexing algorithms.

Discussion

4 comments analyzed.

Competitors mentioned: SQLite with ranges, DuckDB

Concerns raised: DuckDB's vector indexes not portable/shareable as standalone artifacts, DuckDB vector indexes require database server for querying

Competitors

Other products that read as similar to this one — 195 launches clear the similarity bar, closest 8 shown.

Attention rank: #21 of 196 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 294 days after the earliest competitor.

Other launches for this product