krz/brand-bench

 1┌──────────────────────────────────────────────────────────────┐
 2│  B R A N D   B E N C H                     [ KRZ ]   krz.sh  │
 3└──────────────────────────────────────────────────────────────┘
 4
 5WHAT
 6  client-side brand identity generator. describe a project, get a
 7  brand package: positioning, tone, color palette, typography, logo
 8  concepts, taglines, usage examples. runs in the browser.
 9
10  no backend. no accounts. state persists in localStorage.
11
12DOES
13  - template generation, offline, tailored by category
14  - ai generation via a local ollama instance, cancellable mid-stream
15  - multiple packages in tabs, each persisted separately
16  - inline editing, click any field in the preview
17  - section locking across regenerations
18  - undo/redo per package (⌘Z / ⌘⇧Z)
19  - editable color swatches, native color picker
20  - font pairings per category, 8-level type scale
21  - export to markdown, json, or a self-contained html page
22
23RUN
24      npm install
25      npm run dev
26
27  open http://localhost:5173.
28
29  build:
30
31      npm run build      # dist/
32      npm run preview
33
34AI WITH OLLAMA
35  1. install ollama, pull a model:
36
37         ollama pull llama3.2
38         ollama pull mistral
39         ollama pull gemma3
40
41  2. run it:
42
43         ollama serve
44
45  3. a non-default host or port needs OLLAMA_ORIGINS set for browser
46     requests:
47
48         OLLAMA_ORIGINS="*" ollama serve
49
50  4. in the app: Settings, enable AI, set base url
51     (http://localhost:11434), pick a model, Test connection.
52
53  bigger models give better-structured output. generation takes
54  15-60s depending on hardware.
55
56DOCKER
57  Dockerfile and compose.yml deploy the app next to ollama.
58
59      docker compose up -d --build
60
61  app serves on port 8000. ollama's api is proxied through nginx at
62  /api/, so the browser makes no cross-origin request. after first
63  boot, pull a model:
64
65      docker compose exec ollama ollama pull llama3.2
66
67  set the ollama base url to http://your-server:8000. for an nvidia
68  gpu, uncomment the deploy block in compose.yml (needs the nvidia
69  container toolkit).
70
71STACK
72  react 18, typescript, vite. no ui library, plain css.
73
74LICENSE
75  0bsd. see LICENSE.
76
77┌──────────────────────────────────────────────────────────────┐
78│  krz.sh                                                      │
79└──────────────────────────────────────────────────────────────┘