Contora
Start your free trial

Use Claude with Contora

Two optional skills for people who use Claude. One renders your Render Kits into finished photos, as many vases at a time as you like. The other turns a room photo you like into a new scene for your vases.

contora-render-clay-image

Turns your -clay.png and -prompt.txt files into finished listing photos.

Download the render skill

contora-scene-prompt

Turns a room photo into a -prompt.txt scene you can render any vase in.

Download the scene skill

Don't use Claude? You don't need these. The Render Kit guide shows how to make photos in any AI image tool.

Before you start

Install the skills

  1. Download the .zip files

    Use the buttons above. Don't unzip them.

  2. Upload them to Claude

    In Claude, open Customize, then Skills, and upload each .zip. If you don't see Skills, turn on code execution first.

  3. Connect an image generator

    Add your image service under Claude's connectors and sign in to it. You only need this for the render skill.

Calling a skill

Start a new chat, attach your files, and begin your message with the skill's name: type / and pick it from the list, or type it in full, such as /contora-render-clay-image. Claude will usually pick the right skill by itself when you drop in Contora files, but naming it makes sure.

Keep Contora's file names as they are. The render skill matches files by name, so Sprits-clay.png goes with Sprits-prompt.txt.

Render your vases

Export a Render Kit for each vase from Contora (Export, then AI Render). Each gives you a -clay.png and a -prompt.txt. Unless you say otherwise, the skill makes one 1K square image per vase.

1. One vase, one scene

The simplest case: one Render Kit.

Attach
Sprits-clay.png, Sprits-prompt.txt
Type
/contora-render-clay-image Render this vase.
You get
Sprits-render.png

2. One vase, your own settings

Say the size (1K, 2K or 4K), the shape (1:1, 4:5, 3:4, 2:3, 16:9, 9:16) and how many versions (1 to 5).

Attach
Sprits-clay.png, Sprits-prompt.txt
Type
/contora-render-clay-image Render this at 2K in 4:5 portrait. Give me three versions.
You get
Sprits-render-1.png, -2 and -3

3. Several vases, each in its own scene

Attach several Render Kits at once. Each prompt is matched to the clay render with the same name.

Attach
Sprits-clay.png, Sprits-prompt.txt, Otto-clay.png, Otto-prompt.txt, Veron-clay.png, Veron-prompt.txt
Type
/contora-render-clay-image Render all three vases at 2K.
You get
3 photos: Sprits-render.png, Otto-render.png, Veron-render.png

4. One scene, several vases

Attach one prompt and several clay renders, and every vase goes into that scene. Good for a matching set of listing photos.

Attach
Sprits-prompt.txt, Sprits-clay.png, Otto-clay.png, Veron-clay.png
Type
/contora-render-clay-image Put all three vases in this scene.
You get
3 photos, all in the Sprits scene

5. Every vase in every scene

Attach several clay renders and several scene prompts, and ask for every combination. Three vases in two scenes makes six photos.

Attach
Sprits-clay.png, Otto-clay.png, Veron-clay.png, linen-shelf-prompt.txt, marble-kitchen-prompt.txt
Type
/contora-render-clay-image Render every vase in every scene at 2K.
You get
6 photos, named vase then scene: Sprits-linen-shelf-render.png, Sprits-marble-kitchen-render.png, and so on

What Claude shows before it renders

Nothing is generated until you say yes. Claude sends one summary like this:

About to render: Sprits, Otto and Veron, each in Linen shelf and White marble kitchen (6 images)

Settings: 2K, 1:1, 1 image each

Cost: 72 credits (410 left)

Reply yes to run, or change the size, shape or number of images.

Reply yes, or change something first, for example "make it 1K" or "just the marble kitchen". Costs multiply: 3 vases in 2 scenes with 2 versions each is 12 images, so try 1K first on a big batch. When the photos arrive, compare each with its clay render and ask Claude to redo any where the shape changed, for example "run Otto in the linen shelf again".

Make a scene from any photo

Give the scene skill a room or styled photo you like, and it writes a -prompt.txt in the same format as Contora's own. Use that file with your clay renders, as in examples 4 and 5 above, to put your vases in that scene. The skill needs a name for each scene and one vase colour: one of the 8 Contora finishes, or your own colour's name and hex value. If you leave either out, Claude asks.

6. One photo

Attach
oak-shelf-photo.jpg
Type
/contora-scene-prompt Turn this photo into a scene called Oak shelf, in Sage.
You get
oak-shelf-prompt.txt

7. Several photos at once

Each photo becomes its own scene. The colour you give applies to all of them; for different colours, run each colour separately.

Attach
photo-1.jpg, photo-2.jpg, photo-3.jpg
Type
/contora-scene-prompt Make scenes from these three photos in Oxblood. Call them Gallery wall, Reading nook and Dark kitchen.
You get
gallery-wall-prompt.txt, reading-nook-prompt.txt, dark-kitchen-prompt.txt

8. Your own colour

For a colour you created in Contora's Scene Manager, give its name and hex value.

Attach
terrace-photo.jpg
Type
/contora-scene-prompt Make a scene called Terrace from this photo, in my colour Terracotta, #A4593A.
You get
terrace-prompt.txt

Both skills in one chat

You can make new scenes and render your vases in them without leaving the chat. The render skill picks up the scene prompts made earlier in the same conversation.

9. New scenes, then a full set of photos

Attach
oak-photo.jpg, bench-photo.jpg, Sprits-clay.png, Otto-clay.png, Veron-clay.png
Type
/contora-scene-prompt Make scenes from the two photos in Sand, called Oak shelf and Linen bench.
/contora-render-clay-image Now render all three vases in both new scenes at 2K.
You get
Two scene prompts, then 6 photos: each vase in each new scene

Keep your new -prompt.txt files. Drop them into any future chat with new clay renders to reuse the scene.