I embedded 10M StreetView images
Details
- External ID
- 46387007
- Source
- HN
- Company
- —
- Product
- I embedded 10M StreetView images
- Website domain
- sdan.io
- Launched
- Dec. 25, 2025
- Cohort
- —
- Upvotes
- 14
- Upvotes percentile
- 0.5868320610687023
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:25 p.m.
- Updated at
- Sept. 7, 2026, 9:25 p.m.
Enrichment
- Theme
- niche creative and graphics software
- Vertical
- Horizontal
- Function
- Data infrastructure
- Audience
- Developer
- AI stance
- Not AI
- Project type
- Hobby / open-source project
- Normalized one-liner
- streetview image dataset with 10 million images
- Manually corrected
- False
Could you build this?
No Scraping, processing, and generating embeddings for 10 million Street View images requires substantial compute infrastructure, high API/proxy costs, and specialized vector indexing at scale.
What it would actually take: Requires a distributed data ingestion pipeline to query or scrape 10M panoramic images, a large GPU cluster running vision embedding models (e.g., CLIP or SigLIP), and a high-performance vector index (HNSW/Faiss/ScaNN). The browser UI requires custom WebGL/Three.js point cloud rendering with spatial level-of-detail (LOD) tiling to display millions of coordinates smoothly. Needs substantial cloud infrastructure funding, distributed data processing, and large-scale 3D graphics expertise.
Discussion
5 comments analyzed.
Competitors mentioned: Thinking Machines' Tinker API
Concerns raised: Error is quite high, often in the tens of thousands of km, Some images are upside-down, Model prone to overfitting
Feature requests: Ability to win at geoguesser, Distill expert knowledge (like Rainbolt's approach) into model
Competitors
Other products that read as similar to this one — 269 launches clear the similarity bar, closest 8 shown.
Attention rank: #102 of 270 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 54 days after the earliest competitor.
- I compressed videos by upto 78x while preserving faces,text and details · hn · 2026-08-28 · 5 upvotes · similarity 0.49
- I run a vision model on every screenshot, locally, on a 4GB GPU · hn · 2026-06-13 · 36 upvotes · similarity 0.49
- I built 48 lightweight SVG backgrounds you can copy/paste · hn · 2026-03-18 · 403 upvotes · similarity 0.48
- I built a tiny camera that knows where it is · hn · 2026-09-15 · 5 upvotes · similarity 0.46
- iPhone app takes simultaneous images from 2 lenses, fuses into 1 photo · hn · 2026-08-08 · 329 upvotes · similarity 0.45
- farview-but-good · github · 2026-09-16 · 8 upvotes · similarity 0.45
- Rendering 18,000 videos in real-time with Python · hn · 2026-02-22 · 49 upvotes · similarity 0.44
- 1080p is 920px tall · hn · 2026-09-14 · 42 upvotes · similarity 0.44
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a data infrastructure tool for Media & entertainment yet.