We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
Fix: Replace fake shimmyvision.ai URLs with real Payment Links - Get a License now links to Developer tier Payment Link - Renew license links to correct Payment Link - All links properly redirect through workers.dev webhook for license delivery
Add Shimmy Vision documentation - Shimmy-Vision.md: Main overview page with features and licensing - Shimmy-Vision-API.md: Complete API reference with endpoints - Shimmy-Vision-Quick-Start.md: 5-minute getting started guide - Shimmy-Vision-Modes.md: Detailed mode descriptions (ocr, layout, brief, web, full) - Shimmy-Vision-Troubleshooting.md: Common issues and HTTP error codes - Home.md: Added Vision section to navigation
docs: clarify MLX backend development status in wiki TRANSPARENCY UPDATE: Corrects misleading MLX documentation to reflect actual development status, addressing community feedback. 🍎 Apple Metal Section Updated: - Clarified that Metal acceleration comes via llama.cpp Metal backend - Added prominent note explaining MLX backend development status - Emphasized current excellent performance through battle-tested llama.cpp - Set accurate expectations for future native MLX implementation 📝 Development Honesty: - Explained distinction between current Metal acceleration and planned MLX - Clarified that .npz MLX-native model support is coming in future releases - Maintained emphasis on excellent current Apple Silicon performance - Removed any confusion about current MLX implementation status This addresses community feedback about documentation accuracy and ensures users have correct expectations about Apple Silicon acceleration capabilities. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
feat: modernize wiki with MOE documentation and current capabilities CRITICAL UPDATE: Added comprehensive MOE (Mixture of Experts) documentation that was completely missing from the wiki. This addresses the major gap in documenting Shimmy's large model capabilities. 📚 Home.md Updates: - Updated feature list to highlight MOE CPU/GPU offloading - Added smart model filtering and release gate quality mentions - Updated installation commands to include MOE features - Modernized capability descriptions to reflect v1.7.1 features 🖥️ GPU-Support.md Major Enhancement: - Added comprehensive MOE section explaining hybrid processing - Detailed MOE configuration with --cpu-moe and --n-cpu-moe flags - When to use MOE guide for large models and limited VRAM - Updated CUDA installation to include MOE combinations - Added M3 Pro support clarification ⚙️ Installation-Guide.md Updates: - Added MOE feature installation options - Updated cargo install commands with MOE combinations - Recommended installation now includes MOE features - Added all-features installation option 🔗 Navigation Improvements: - Removed broken links to non-existent documentation pages - Added "coming soon" status for planned documentation - Focused on existing, accurate content This brings the wiki documentation in line with Shimmy's actual capabilities and ensures users understand the MOE features that enable running large models on consumer hardware. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
feat: Add comprehensive wiki documentation - Home page with navigation - Complete installation guide for all platforms - Detailed GPU support documentation with CUDA/Metal setup - Performance comparisons and troubleshooting guides
Initial Home page