Nicheloom

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

Rampart on CoreML

Details

External ID
48733651
Source
HN
Company
—
Product
Rampart on CoreML
Website domain
github.com
Launched
June 30, 2026
Cohort
—
Upvotes
6
Upvotes percentile
0.31420765027322406
Tags
—
Fetched at
Sept. 7, 2026, 9:26 p.m.
Updated at
Sept. 7, 2026, 9:26 p.m.

Description

CoreML implementation of National Design Studio's Rampart PII detection model + Swift package

Enrichment

Theme
gpu compute and acceleration tools
Vertical
Security
Function
Model & infra
Audience
Developer
AI stance
AI feature
Project type
Commercial product
Normalized one-liner
rampart on coreml
Manually corrected
False

Could you build this?

Partial Creating a Swift wrapper is straightforward, but converting a machine learning model to Apple's CoreML format while maintaining accuracy, tokenization parity, and hardware acceleration requires ML conversion expertise.

What it would actually take: Requires using Apple's coremltools to convert PyTorch/HuggingFace tokenizers and Transformer backbones into .mlpackage format, managing custom ops and quantization (e.g., int8/int4 weights via MIL). The hardest aspect is porting custom subword tokenizers to pure Swift or CoreML-compatible layers and ensuring tensor operations execute cleanly across Apple Neural Engine (ANE) without fallback CPU bottlenecks.

Discussion

2 comments analyzed.

Concerns raised: High LLM costs for entity stripping

Competitors

Other products that read as similar to this one — 922 launches clear the similarity bar, closest 8 shown.

Attention rank: #607 of 923 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 242 days after the earliest competitor.

Other launches for this product

Same idea, different domain

Nobody's really built a model & infra tool for Fintech yet.