What it is in one line
A personal project that uses photogrammetry to move his local climbing gym into the browser: about ten minutes of iPhone 14 filming becomes a draggable 3D gym, with routes marked on every wall, climbing videos pinned back to the exact spot they happened, and first-person walking inside the building. Its thesis: the right way to record climbing is neither photos nor video, but 3D.
Who built it
GitHub user kmcheung12, a personal project hosted on GitHub Pages. The repository was created 2026-07-23; the author's full name and background are not disclosed on the page.
Read: this is a "the gym down my street" personal project — not built as a general product, just to solve his own "how do I record climbing" pain, and it reads as true precisely because it is specific.
What it actually does
- 3D gym tour → drag to look around the whole gym, including the overhang island and the slab island, with first-person WASD walking
- Route annotations → start/finish holds and hold counts (e.g., "Route 2 has 19 holds") marked directly on the walls
- Climbing videos pinned back to the wall → attempts that used to rot in WhatsApp or Instagram are placed back where they physically happened
- Walk around the climber → the person in the video becomes a 3D body you can orbit, revealing hip rotations a head-on camera could not show
What old behavior it replaces
Two things.
First, how gyms and routes get recorded and shared: photos flatten the wall, video fixes you at the camera's viewpoint — neither conveys the terrain, and both lock you to one angle. Replaced by an orbitable 3D space.
Second, how climbers preview problems: climbers read routes standing at the wall (deciding hand sequence in advance) or scroll social media for line previews. This project moves route-reading onto a webpage — you can work out the sequence in 3D before you even go to the gym.
Read: for climbing, a route is information in 3D space; photos and video each lose a dimension. This project adds back exactly that dimension.
Business model
None. A free personal project with no product ambitions.
Read: projects like this usually stay at the "artwork" level, but its value is validating a format: 3D reconstruction of a physical space + information annotation + first-person walkthrough. The combination transfers to gyms, showrooms, shops, venues — anywhere the space itself is the product.
Hard numbers
- HN: 28 points, 11 comments
- Capture cost: iPhone 14, about 2 minutes per wall, about 10 minutes total
- GitHub repo: 0 stars, created 2026-07-23
- Example covers 5 walls, 2 islands, several annotated routes, and one wall-pinned video
- Author information, any business plan: not disclosed
Four-way read
| Dimension |
Call |
| Founder-product fit |
The author is the climber himself; the pain (recording climbing in 3D) is real and self-served |
| Product insight |
"Pin media back to the 3D position where it happened" is the most valuable use of photogrammetry content |
| Execution quality |
The result is smooth, navigable, annotated — not a crude toy, though it leans on existing toolchains |
| Timing |
Phone photogrammetry is mature and the spatial-computing narrative is heating up — right when "spatial content" is being repriced |
The call
The strongest asset of a personal project is specificity. "The climbing gym down my street" beats "photogrammetry technology" by a mile — the 28 HN points bought that specificity. This project has no product ambitions, yet it performed the best product demo possible: letting readers immediately see what new information "3D-ifying a physical space" produces.
The transferable pattern: give spatial content a narrative layer. A raw 3D scan is data; add route annotations, first-person narration, and video pinning and the data becomes consumable content. Any team digitizing spaces, venues, or physical products can copy the "scan + annotate + tour" combination.
The limit: technically there is nothing new — iPhone photogrammetry plus web rendering. The value sits entirely in the format and the specificity. To become a product it would need large-scale capture, updating, and rights handling, none of which are touched.
What to watch next
① Whether the author makes a second space (continued output = the format validating itself)
② Whether anyone turns it into a tool or service for the climbing-gym industry — commercialization is the format getting validated
③ Whether climbing-industry commenters signal a real need in the thread
What you can take from it
Content logic: for showcase projects, pick a specific object in the reader's world, not abstract demo data. Specificity travels because readers can map it onto their own lives.
Format template: the "3D scan + annotations + first-person tour + media pinning" combination is directly reusable — for any physical space (retail, showroom, studio, gym) it is a ready-made content and product template.
Verdict
Worth watching — as a format, not a product. No new technology, no product intent, but it demonstrates the correct content shape for spatial 3D-ification clearly. For anyone doing AI-driven creation or space digitization, this is a free format template.