Audio-driven long-form output with thumbnail, metadata, and archive-ready manifest.
Real render — widescreen output, captions, and publish metadata from an actual Cast job.
Cast Gallery (powered by E3D)Real and representative output from each preset — see what a render actually looks like before you spend credits.
Audio-driven long-form output with thumbnail, metadata, and archive-ready manifest.
Real render — widescreen output, captions, and publish metadata from an actual Cast job.
Real ~9-minute podcast rendered end to end with AssemblyAI diarization and Pexels B-roll.
Real render — 70 diarized segments, semantically-matched B-roll, burned captions, 8m59s.
Submitted the way an agent would: a source audio URL and a curl-shaped API call, preset "short" with the finance_signal style.
Real render — AssemblyAI diarization, semantically-matched B-roll, burned captions, 40s.
Type one sentence and AI writes and renders the video — birthday and congratulations video cards with two voices and optional email delivery, or a full podcast video from a transcript or audio file.
Promo short about Cast itself, made with the same TTS/B-roll/caption engine that powers every Cast render — not a job submitted through the paid API.
Text-only input, no source audio: a written transcript explaining codex-spec-runner, an open-source CLI that runs Codex CLI or Claude CLI phase-by-phase from a Markdown feature spec, with model routing and resumable execution.
Real render — synthesized narration, semantically-matched B-roll, burned captions, 57s. View the repo ↗
Want to make one of these? Head back to the Cast workspace — try a free sample, or buy the Starter pack for $9 with a card (no wallet required) and create a real video.