Rebuild a real place from several photos, a panorama or a phone capture
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
A world at metric scale with the floor level and the up axis stated, returned as splat and mesh files through the API, on a plan the agent's account can use without a sales contract.
Measured, for our own tests: the world matches the real place from the input viewpoints. No docs we read state these figures, so they don't count against a listing here.
How agents ask for it: make a 3D model of this room from these photos
, turn my phone video of the workshop into a 3D space
, build the splat from all these angles
Does the job, with a limit 2
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| SpAItial API (Echo) C SpAItial Ltd | Echo-3 takes a capture_bundle ZIP (frames, per-frame intrinsics, optional poses) at real-world scale, self-serve from Standard, 900 credits plus 1 a frame, 30 to 120 minutes. Plain photos or a video aren't accepted, the floor level isn't stated, and commercial use needs Pro. source |
| World Labs World API (Marble) D World Labs Technologies, Inc. | Self-serve multi-image (4 images, 8 with reconstruct_images) and video (100 MB) inputs return splats, collider and HQ mesh with metric_scale_factor and ground_plane_offset. A factor of 1.0 means scale wasn't inferred, the up axis is OpenCV (+y down), and match to the real place is unmeasured. source |
What stops agents finishing
- phone-capture reconstruction is enterprise-only on the SpAItial API
- single-image tools can't take several photos of the room (acc-11)
- terms for showing captured or gallery worlds unconfirmed (acc-18)
- worlds may be public unless the request says otherwise (acc-16)
Around it
Before: file-storage.read-file. After: 3d-worlds.load-walkable, 3d-worlds.sim-environment.
Where this comes from
Weight 17 of 100 in 3D worlds: 6 of 36 links: GitHub issues asking for multi-image input or using Echo-3 and multi-image Marble jobs, and one robot-simulation plan
The same as data: /api/v1/use-cases/3d-worlds.capture-to-world.json, or the MCP tool get_use_case with id 3d-worlds.capture-to-world. Other jobs in this category: /api/v1/niches?category=3d-worlds.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.