lcmind/vibe-link-server

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

fix: S-tier prompt - hex to color name, blur UI text, remove watermarks

feat: Flux generates brand text directly - remove overlay, prompt master typography rules

refactor: 10-year prompt engineer redesign - Flux.1-schnell optimal params + smart text overlay

feat: add Pillow brand overlay + fix prompts to show actual business visuals (not random space scenes)

refactor: scene-first approach - focus on visual metaphor without text generation

refactor: revolutionary prompt - analyze WHAT they DO, not just category templates

refactor: apply Legendary Marketing CEO approach with TEXT IS HERO and business metaphors

refactor: apply Minimalist Brand Director prompt with Hero Object and category-based styling

feat: switch to 9:16 vertical format with optimized Flux settings (768x1344, steps=28, guidance=3.5)

refactor: improve Gemini and Flux prompts with structured format

feat: add CORS support for localhost:3001 and 3002

fix: correct router.huggingface.co URL path

fix: use correct HF router endpoint URL

fix: use direct httpx API call for HF inference

fix: update HF InferenceClient to use new router endpoint

fix: switch to gemini-2.5-flash

debug: add detailed error logging to /api/create

fix: correct pydantic_settings usage with Field aliases

fix: switch from test_app to real app.main:app

fix: simplify Dockerfile to minimal HF Spaces setup

test: add simple test app to verify HF Spaces setup

fix: add PYTHONPATH for module imports

fix: follow HF Spaces official pattern - USER after pip install

fix: use python -m uvicorn and install deps as user

fix: add HF Spaces required user (UID 1000) and proper permissions

fix: add proxy headers support for HF Spaces

fix: switch to Docker SDK with direct FastAPI, remove Gradio wrapper

fix: update to Python 3.11 and Gradio 6.3.0 for HF Spaces compatibility

refactor: S-grade architecture with layered structure and strict CORS

fix: add app_port to README metadata for HF Spaces routing

fix: optimize uvicorn for HF Spaces proxy (add proxy-headers and forwarded-allow-ips)

fix: add root-path for HF Spaces proxy compatibility

fix: restore requirements.txt for Docker build

fix: revert to requirements.txt for simpler build

fix: correct colorFrom value for HF Spaces

docs: add HF Spaces config metadata to README