Mira-Scene
A solid pick for self-hosting and full data control.
Overview
Mira-Scene is a 3D scene reconstruction framework released in September 2026 by VAST (the team behind TripoSG) with the University of Hong Kong. From a single photo, it separates each object, estimates depth, generates a mesh per object and places it back in the scene at the right position and scale, with an environment map for lighting. The output is a scene you can edit object by object, not a frozen 3D block. Its technical contribution targets object placement, the weak spot of compositional methods. Instead of regressing one pose per object, the model generates, for every visible pixel, a coordinate in the object's own space, then aligns those correspondences with the estimated depth to recover position, orientation and scale. The authors report relative gains of 39.8% in 3D-IoU and 16.5% in 2D-IoU over SAM3D, trained on limited open data. Code, datasets and weights (pretraining and final) are released under MIT on GitHub and Hugging Face. It is a multi-stage research pipeline, not a ready-made app: it needs an NVIDIA GPU and a Python setup, and the vendor publishes no memory requirement. The bundled web viewer listens on all interfaces without authentication, so do not expose it on an untrusted network. For a room, a set or a game scene it is a useful building block; for a single object to 3D-print, an object-level generator such as TRELLIS.2 remains more direct.
Skill profile
Not disclosed
Strengths
- Has a free tier
- Open-source and self-hostable
Who is it for
- you want to control cost or self-host
- your data must stay in the EU
Ideal use cases
- Rebuilding a room or set in 3D from a photo
- Preparing scenes for architectural visualisation or games
- Research on compositional 3D reconstruction
Access & availability
Key specifications
What it costs per month
Estimate from sourced API prices and a hardware tier for self-hosting. Assumptions are editable.
Assumptions: 36-month amortisation, 0.25 €/kWh, 22 days/month, 1 $ = 0.92 €. Excludes labour, software licences, redundancy.
nAIvigate Trust Index
What the vendor offers and documents for professional use in Europe. An undocumented criterion counts as zero and stays visible: the score rises when the vendor publishes, not when we assume.
19/19 criteria documented · 100 % · 12 of them set by rule
See the criteria▾
Trust Index badge
Show this score in your docs, README or security page. It updates automatically.
[](https://naivigate.com/en/model/mira-scene)Data & compliance
What we know, line by line, with source and date. These lines feed the Data and Enterprise families of the Trust Index above. GDPR compliance depends on your processing, not on the model alone.
| EU data residency | Yes |
| Data processing agreement (DPA) | N/A· self-hosted: data never leaves |
| Transfer outside the EU | No |
| Training on your data | No |
| Configurable retention | N/A· self-hosted: data never leaves |
| Zero Data Retention | N/A· self-hosted: data never leaves |
| Self-hostable | Yes |
| Declared hosting | self |
Quality of this card’s data
This score measures the quality of the data on this card: price, context, Elo, licence, documented, sourced and fresh. It measures neither the model nor its compliance: for that, see the Trust Index above.
2/8 documented · 1/8 primary sources
Run it locally
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve Yang-Tian/Mira-Scene \
--max-model-len 32768 \
--tensor-parallel-size 1 \
--dtype autoAdvanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |