bestblogs-podcast-video

What problem does it solve?

This Skill automates converting a BestBlogs daily digest into a production-ready podcast audio file and a synchronized short video, removing manual steps like script drafting, TTS slicing, image curation, audio merging, video rendering, and upload. It streamlines turning the daily 10-item digest into a branded 10–12 minute podcast and an accompanying MP4 video with consistent visuals.

Core Features & Use Cases

• Script generation for a Top 3 deep-dive + 7-item quick-review format with user confirmation before synthesis.
• Segmented TTS synthesis using Fish.audio with retry and merge logic, FFmpeg loudness normalization, and optional voice model override.
• Image sourcing, quality checks and fallback image generation via the project's image-gen skill to ensure consistent brand visuals.
• Remotion-based video data generation and rendering pipeline that aligns per-segment audio timestamps to scenes, plus utilities to copy assets into Remotion public/ for rendering.
• R2 upload and RSS feed generation/update with metadata.json and podcast.xml support, including safe overwrite and caching strategies.
• Use cases: daily automated episode production, regenerate audio/video with different voice models, or generate audio-only episodes without uploading.

Quick Start

Generate today's BestBlogs podcast and short video using my cloned Fish.audio voice, then upload the episode and RSS update to R2.