shadertoy-cli 2.1.5

Agent-friendly ShaderToy project, rendering, debugging, and live-preview CLI
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44

ShaderToy CLI agent workflow

To start from an existing public ShaderToy:

     shadertoy import https://www.shadertoy.com/view/XXXXXX -o project

This creates the same editable directory format used by shadertoy new and keeps
supported remote assets local.

1. Start by inspecting the project:
     shadertoy inspect --json

2. Edit ShaderToy.toml and files under shaders/ or assets/.
   The manifest is schema-backed. To print the exact schema:
     shadertoy docs manifest --schema

3. Validate after edits:
     shadertoy check --json

4. Render deterministic evidence:
     shadertoy render -o target/check.png
   For temporal comparison/contact sheets, reuse one runtime:
     shadertoy render-frames --frames 0,60,120,180 --contact-sheet target/contact.png

5. Debug multipass projects from the outside in:
     shadertoy inspect graph --json
     shadertoy inspect pass buffer-a --json
     shadertoy render --pass buffer-a -o target/buffer-a.png

6. Freeze a feedback state when a bug appears:
     shadertoy state capture --frame 300 -o target/frame300.ststate
     shadertoy state inspect target/frame300.ststate --json

7. Replace a buffer with a known image to isolate a pass:
     shadertoy render --state target/frame300.ststate \
       --set-buffer buffer-a=fixtures/a.png \
       -o target/debug.png

8. Use live native-rendered review when iterating:
     shadertoy preview

On Linux, check/render/render-frames/state capture/preview use surfaceless EGL and do not require DISPLAY or WAYLAND_DISPLAY.

Do not edit target/. It is disposable generated output.