Analysis and planning
AI Room Analysis
The scored critique: every field it returns, how to read the scores, and what it rejects.
AI Room Analysis reads a photo of a real room and returns a structured critique instead of a redesign. It requires Pro and does not consume a design generation.
Response fields
| Field | Type | Meaning |
|---|---|---|
| Room type | Text | What the AI identified the space as |
| Overall score | 0–100 | Headline judgement of the room as it is |
| Architecture score | 0–100 | Proportion, structure, light and organisation |
| Architecture critique | Text | Written explanation of the architecture score |
| Wellbeing score | 0–100 | How the room feels to occupy |
| Wellbeing observations | 3 items | Light, air, flow or calm |
| Strengths | 3 items | What works and should be kept |
| Weaknesses | 3 items | Specific problems worth fixing |
| Improvement prompt | Text | An instruction you can hand to the AI Interior Designer |
Reading the scores
The scores are a consistent AI judgement, not a measurement. Their value is comparative: analyse the room, generate a redesign, analyse the redesign, and read the difference. Small variations between runs on the same photo are normal; large swings usually mean the photo is ambiguous.
Rejections
Inputs that are not photographs of a real space — people, product shots, documents, screenshots, sketches, renders, or images too dark or too close to read — are refused with a "not a room" message. This is deliberate: scores for a non-room would be meaningless.
Language
The analysis is written in your app language across all 23 supported languages.
Common questions about this
Related reference
Ready to try it on your own room?
Three AI designs every day — no account and no payment card required.