Content Automation & Story Tool — turns Reddit-style stories into narrated short-form video, one at a time or in bulk.
CAST writes a story, narrates it with text-to-speech, generates burned-in captions, and renders it over a background clip — the same format used across short-form video platforms. It runs entirely in your browser (or as a standalone desktop app), with no account required and no data sent to any server the developer controls.
Story text is generated by an AI provider you configure with your own API key (or a local model via Ollama, for free). Narration runs client-side using open TTS engines, or an optional cloud voice provider if you supply a key for one. Everything — your API keys, your generated videos, your connected accounts — stays on your own device.
DeepSeek, OpenAI, Groq, OpenRouter, Mistral, Gemini, Anthropic, or a local Ollama model. Bring your own key, or run Ollama for free.
Several TTS engines with adjustable speed, stability, and pitch — free in-browser options, or paid cloud voices if you prefer.
Burned-in captions with real style presets — word-by-word, grouped phrases, or karaoke-style highlighting, with entrance animation, background boxes, and shadows.
Generate a whole batch of videos in one pass from a library of your own background clips and music.
Optionally connect a YouTube channel to auto-generate titles, descriptions, and thumbnails, and schedule uploads directly from the app.
Works with YouTubeRunning locally with ffmpeg on your PATH gives faster, more reliable rendering than the in-browser fallback.
Runs entirely in your browser. Slower on longer videos than the native app, but nothing to install.
Faster native rendering. Unsigned — macOS will show a Gatekeeper warning on first launch (right-click → Open to bypass it, since there's no Apple developer certificate behind this build).
Windows and Linux builds aren't published yet — you can build them yourself from source; see the README.