# Lamp Layers

> A lighting planner for one room. Place the room's lamps (up to ten: ceiling light, pendant,
> downlight, shaded floor lamp, table lamp, angled reading lamp, desk lamp, under-cabinet strip,
> wall uplight, picture light, hidden LED strip) and the spots where things happen (up to four:
> reading, relaxing or TV, desk work, dining, kitchen counter, bedside) on a to-scale plan. A free
> in-browser engine estimates the light each spot gets from each lamp, compares it with the target
> for that activity, reads the ambient, task and accent layers and names a verdict: short of light,
> a layer missing or ready for scenes. Dimmer scenes show the light at every spot. A metered lighting
> plan writes one dimmer scene per spot, the lamps worth adding and a check to make first. Derived
> from @curiositech/interior-design-expert.

https://lamp-layers.skillsafe.ai/

## What it does

- The estimate: each lamp type has a typical height and a share of its light aimed down; that share spreads and fades with the square of the distance, and the rest bounces off walls and ceiling and arrives everywhere (more with light walls). The page adds every lamp's share at each spot at table height. Daylight is not counted. Estimates, not measurements.
- Targets by activity: reading and desk work 300-500 lux, relaxing and dining 150-300, kitchen counter 300-750, bedside 50-150. Colour temperature by time of day: morning 5000-6500 K, midday 4000-5000 K, evening 2700-3000 K, night 2200-2700 K.
- Verdict, checked in order: short of light when a spot stays under its target with every lamp at full; otherwise a layer missing when there is no ambient lamp, no accent lamp, or no task lamp while the room has a reading, desk or counter spot; otherwise ready for scenes.
- Drag lamps and spots on the plan; the glow and the readouts update as you go. Make dimmer scenes (0-100% in tens; a lamp that cannot dim is off or full) and see each spot against its target.
- Free downloads: the plan (SVG, PNG), the spots with each lamp's share (CSV), and everything with the plan (Markdown, JSON).
- Saved rooms (opt-in) keep the lamps, spots, scenes and the latest plan, and show how the verdict moved between runs.

## The paid task

One task, `light`, on gpt-terra, metered per run. The input is five string fields: `task`, `room` (kind, size, ceiling, walls, name, note), `lamps` (one a line), `spots` (one a line) and `facts` (the page's estimates, layers and verdict). The reply is one JSON object: verdict (short-of-light | missing-a-layer | ready-for-scenes, always the page's), headline, layers_note, scenes (one per spot with a level for every lamp and the light at the spot), additions (catalogue lamps with lumens, colour temperature and the spot they go near) and check_first. The page re-adds every scene's light from its own shares, checks it against the target, checks every level and every added lamp, and places added lamps on the plan to show what they would give.

A planning aid, not professional advice. New fixed lights and wiring are a job for a qualified electrician.

- API tutorial: https://lamp-layers.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@curiositech/interior-design-expert (curiositech/some_claude_skills, MIT)