Description
Knowledge videos take a long chain: pick a topic, research, write the script, record narration, build visuals and check nothing is wrong. vibe-knowledge is a Claude Code skill that runs the chain inside a conversation: give it one fact and it produces a topic, fact list, narration and voiceover, then renders a 1080p60 film frame by frame in code with every line sourced, runs quality checks and packages the release. Three human checkpoints keep you in charge of the key decisions.
It ships with 44 visual styles in 8 families, from classical Chinese and ink to caption news, blueprints and film title cards, and recommends three suited styles after you pick a topic. Most are pure code; a few marked enhanced need an outside image service. A default BGM is mixed so speech stays clear. It pairs with the cine engine from ai-teardown and is MIT licensed.
Seven stage flow: topic, fact list, narration, voiceover, code rendering, QA and release pack, with three human checkpoints.
Every line sourced: facts in the narration carry sources for checking.
44 styles: families such as classical, dark concept, game, documentary, diagram, print and cinematic, with sample reels.
Style recommendation: three styles suggested per topic, or start from a blank custom style.
Audio: a default AI voice plus mandatory BGM mixed under speech.
Enhanced tier: a few styles can use an external image service for art and motion shots.
It ships with 44 visual styles in 8 families, from classical Chinese and ink to caption news, blueprints and film title cards, and recommends three suited styles after you pick a topic. Most are pure code; a few marked enhanced need an outside image service. A default BGM is mixed so speech stays clear. It pairs with the cine engine from ai-teardown and is MIT licensed.
Features
Seven stage flow: topic, fact list, narration, voiceover, code rendering, QA and release pack, with three human checkpoints.
Every line sourced: facts in the narration carry sources for checking.
44 styles: families such as classical, dark concept, game, documentary, diagram, print and cinematic, with sample reels.
Style recommendation: three styles suggested per topic, or start from a blank custom style.
Audio: a default AI voice plus mandatory BGM mixed under speech.
Enhanced tier: a few styles can use an external image service for art and motion shots.
