Nicheloom visitors: 13,662 all-time · 8,824 in the last 7 days

The opportunity tracker for new startups.

Today: 2,299 launchs · Niche of the day: document and PDF processing tools · See full report →

iris-3b-metal

Iris-3B pixel-space text-to-image on Apple Silicon GPUs: Metal-optimized engine, native macOS app, CLI, Python and local HTTP API

Get picks like this daily. The day's top launches, AI/tech news, and a weekly opportunity spotlight — straight to your inbox.

This is 1 of 171 launches in edge device utilities and mobile modding — see how it stacks up on momentum and crowding →

432 other launches read as similar to this one →

Details

External ID
1412377741
Source
GITHUB
Company
—
Product
iris-3b-metal
Website domain
github.com
Launched
Oct. 9, 2026
Cohort
—
Upvotes
45
Upvotes percentile
0.69
Tags
apple-silicon, depth-estimation, diffusion, image-generation, iris-3b, macos, metal, mps, pytorch, super-resolution, swiftui, text-to-image

Enrichment

Niche
edge device utilities and mobile modding
Vertical
Horizontal
Function
Model & infra
Audience
Developer
AI stance
AI-native
Project type
Hobby / open-source project
Normalized one-liner
metal-optimized local text-to-image engine for apple silicon
Manually corrected
False

Could you build this?

No Writing a custom low-level Metal inference engine on Apple Silicon GPUs for a 3B diffusion model requires deep systems programming and GPU shader expertise.

What it would actually take: This requires C++/Objective-C/Swift systems programming utilizing Apple's Metal Performance Shaders (MPS) and custom MSL (Metal Shading Language) compute kernels to optimize memory bandwidth and execution for unified memory architecture. The hard part is manual memory management, custom quantization kernels, and translating PyTorch/safetensors graph execution into highly tuned Metal shaders without framework overhead. It demands high-performance GPU compute and systems ML engineering talent.

Competitors

Other products that read as similar to this one — 432 launches clear the similarity bar, closest 8 shown.

Attention rank: #181 of 433 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 336 days after the earliest competitor.

Other launches for this product