Same flow whether your source is a podcast, an interview, a long YouTube video, or a livestream VOD.
Sign in to AlcheClip, click New Job, and paste any YouTube URL — or drop a video file up to 500MB. The MVP supports source videos up to 15 minutes long. Podcasts, interviews, vlogs, lectures, livestream VODs all work.
AlcheClip uses OpenAI Whisper for word-level transcription, then scores every spoken segment for hook strength using length, filler-word density, and hook-keyword signals. Non-maximum suppression dedupes overlapping candidates, and the 5–12 strongest moments are selected automatically.
Each clip is cropped to 9:16, upscaled to 1080p with Lanczos, sharpened with an unsharp mask, denoised with hqdn3d, color-graded for mobile viewing (contrast 1.08, saturation 1.12, gamma 0.98), and burned with karaoke-style word-level captions generated from the Whisper timestamps.
AlcheClip delivers each clip as an individual H.264 MP4 plus a single zip. Drop them straight into TikTok, Instagram Reels, or YouTube Shorts — captions are burned into the pixels, so no editing app is required on your audience's side.
Most 30-minute podcasts finish in 2–4 minutes of wall-clock time. Longer sources or queue load can reach about 10 minutes. Every step of the pipeline streams live via Supabase Realtime so you can watch progress.
No. AlcheClip burns word-level captions directly into the video pixels and exports a TikTok-ready 9:16 1080p MP4. Drag the file into the upload form on TikTok, Reels, or YouTube Shorts and you're done.
AlcheClip's free tier gives you one full job per month with no credit card. That includes word-level captions, 1080p 9:16 output, and the same enhancement pass as the paid tiers. The fastest free Opus Clips alternative.
Because the entire pipeline was designed around output fidelity rather than clip count. Lanczos upscaling instead of bilinear, mobile-tuned color grading on every clip, explicit unsharp masking and denoising, and word-level caption rendering at the pixel level. No filter is optional; no clip skips the enhancement pass.