Move Explanation to midpoint, bigger font, beep 15%

Fix countdown centering, bigger circle, quieter beep, add Explanation heading

Add 3-2-1 countdown timer with beep sounds during thinking phase

Add hook padding, golden explain color, auto-sizing centered question text

Improve quiz design: hook bg box #ab0042, larger explain text, remove sparkle

Add Quiz Reel module with TTS, frame generation, and video composition

Remove art_reels module completely - not working properly

Add AI Drawing Animator with Groq + fallback templates, update UI

Fix: Groq API key check both GROQ_API and GROQ_API_KEY, improve fallback with varied poses

Fix: Correct chunk duration key, add detailed scene/chunk logging

Fix: Follow exact story_reels workflow - create_captions, SRTParser, FFmpegUtils

Fix TTS call (async), add HF Hub cloud upload for art_reels

Fix: Link TTS and Whisper from app.state instead of reinitializing

Add word-by-word captions to stick figure video like story_reels

Major upgrade: Professional stick figure with multi-character, backgrounds, title screens, thought bubbles, visual metaphors, styled captions

Upgrade Art Reels: AI stick figure with TTS, remove Bangla, use openai/gpt-oss-120b model

Fix Unicode encoding error - Replace checkmark with ASCII text

Add Art Reels module - Minecraft Block Art, Code Drawing, Stick Figure Motivation

Use Google Trends RSS for Trending Now - more reliable than pytrends API

Fix circular import in trends module - create TrendsClient directly in router

Add Trends module with PyTrends - Trending Now, Keyword Research, YouTube Trends

Update Groq model to openai/gpt-oss-120b and Gemini to 2.0-flash

Major Story Reels update: Direct script mode, sticky animation style, improved scene continuity prompts

Update text layout: Heading at 40%, Fact text at 50% from top

Add Pixabay video API: dual search (Pexels + Pixabay), optional both APIs, 9:16 portrait filtering

Fix Pexels: pass full keywords list instead of single keyword to find_video

Increase side padding to 80px for better text margin

Major layout update: heading 80px uppercase at top 15%, fact text 60px bold, 80px gap

Update font sizes: heading 65px, fact text 53px, 4 words per line

Add heading with background support: centered layout, heading BG, fact text outline/shadow

Pexels: select first relevant video/photo instead of random for better accuracy

Pexels: select FIRST most relevant photo instead of random

Fix Pexels import path and simplify image prompt

Simplify image prompt: remove confusing keywords, clear NO TEXT instruction

Fix HF storage: correct import (get_hf_storage) and upload_video args

Fix fact_creator: correct API signatures (NVIDIA/CF/Pexels), ensure 9:16 video output

Add Fact-Image module: generate videos with fact text overlay on AI images

Update log message to reflect current video settings (no crossfade, no zoom)

Fix black screen between images: remove crossfadein effect completely

Fix video playback: strict 2s duration, chain method (no overlap), audio duration cap, and ffmpeg browser compatibility (yuv420p)

Fix NameError: use constant CROSSFADE_DURATION instead of undefined variable

Fix video playback: Add ffmpeg params (yuv420p/faststart), cleanup image composition, add imageio-ffmpeg

Ensure ?download=true in router redirect for cloud URLs

Fix: Add ?download=true to HF URLs for direct download

Fix: get_status returns ready for cloud-uploaded videos (check .cloud file)

Complete cloud storage: return cloud URL when HF upload successful, redirect download

Make HF storage repo private by default for security

Add optional HF Hub cloud storage for videos with auto folders per module

Fix: get_video_path always returns Path (matching video_creator pattern)

Fix download: redirect to /videos static URL for streaming support