chore: Set Kimi-K2-Instruct as default model

feat: Improve streaming UX and fix showwaves syntax

feat: Stream full model output to textbox

feat: Add streaming status updates during command generation

feat: Add GLM-4.7-Flash model, fix Gradio version, add CLAUDE.md

feat: SEO optimization - expanded tags, description, and use cases

revert: Disable cached examples

feat: Enable lazy cached examples (Gradio 6)

feat: Upgrade to Gradio 6.2.0

revert: Disable cached examples due to Gradio compatibility issue

feat: Enable lazy cached examples for faster repeat runs

fix: Add Kimi-K2 model mention to description

fix: Add upload step to description

fix: Replace em dashes with colons/commas

fix: Simplify copy, remove 'no code' messaging

feat: Optimize page content for SEO - FFMPEG in plain English

feat: Switch to Kimi-K2-Instruct model

fix: Add z-index to preview thumbnails to prevent overlap

feat: Switch to GLM-4.7 model with improved prompts

fix: Install mediagallery at runtime for HF Spaces

fix: Add Dockerfile for HF Spaces build, prevent table row hover shift

fix: Prevent table row hover shift in Examples

fix: Include built frontend templates for HF Spaces deployment

feat: Add MediaGallery custom component for mixed media assets

fix: Clarify compose_video function documentation to specify it generates FFmpeg commands, not AI-generated videos

Update app.py

Update app.py

Update app.py

feat: Update model configuration for DeepSeek-V3 with new base URL and model name

feat: Update slideshow error guidance and standard resolution recommendations for better user clarity

feat: Enhance slideshow error guidance and increase command attempts for multiple images

fix: Remove reference to Qwen2.5-Coder in the AI Video Composer description

feat: Enhance get_completion and update functions with error handling and command display for FFMPEG execution

Revert "feat: Implement GPU optimization for FFmpeg commands and enhance filter conversion"

feat: Implement GPU optimization for FFmpeg commands and enhance filter conversion

feat: Enhance video composition functionality with detailed docstring and improved command handling

feat: Add GPU-accelerated FFmpeg command execution and refactor command handling

upgrade gradio to 5.34.1 in readme (#19)

fix: Update gradio version to 5.34.1 in requirements.txt

Update requirements.txt

feat: Update model choice to deepseek-ai/DeepSeek-V3 and remove Qwen model

refactor: Remove Qwen model and update default model choice in examples

fix: Remove undefined content variable and simplify generated command display

refactor: Enhance FFMPEG command generation with step-by-step thinking and improved command extraction

refactor: Reorder model configuration for improved clarity and update model selection logic

refactor: Improve model configuration formatting and ensure consistent API key retrieval

refactor: Use second model in example if available, fallback to first

refactor: Replace hardcoded model names with dynamic model selection

refactor: Update model selection to use MODELS dictionary keys

refactor: centralize model configuration and improve management