From clip to caption, in 5 steps.

Launch CapForge and drag any video file onto the window: .mp4, .mov, .mkv, .webm, or anything ffmpeg understands. You can also use ⌘+O (macOS) or Ctrl+O (Windows) to browse.

The file is decoded locally. Nothing is copied, uploaded, or cached outside your working directory.

• Max file size: whatever your disk can hold
• Vertical, square, or horizontal. All fine
• Longer clips just take longer; nothing to configure

Transcription starts automatically. Wispr reads the audio track, runs an on-device Whisper model, and produces word-level timestamps, not just lines.

Apple Silicon users see near-realtime; on x86, expect roughly 1× realtime for English, up to 2× for less common languages.

• Auto-detects the language, or pick one manually
• 99+ supported languages
• Click any word to replay that exact moment

Six motion presets ship in the box. Each one is a carefully-tuned starting point, not a locked template. Click to apply; every property is editable in the next step.

• Highlight: underline/box that follows the current word
• Karaoke Fill: color wipes left-to-right as the word is spoken
• Bounce: each word pops in with a small vertical spring
• Reveal: clean slide-up reveal, one word at a time
• Script: handwritten-feel words, softer timing
• Chunky: large, high-contrast, mobile-first

The right rail exposes every property of your caption: typography, color, shadow, layout, timing, and per-word animation. Drag a slider, the preview updates on the next frame.

Want your brand font? Drop a .ttf or .otf into the typography section. Want per-word color swaps? Turn on active color and the current word will use it.

• Undo / redo everything, forever
• Save your settings as a custom preset
• All values are just numbers; copy/paste between projects

Happy with the look? Hit ⌘+E (or Ctrl+E) to open the exporter. Pick a format and your clip renders with hardware acceleration when available.

Three flavors of output:

• Burned-in video: MP4 or MOV, captions welded to every frame
• Transparent overlay: MOV with alpha; drop into your NLE as a separate track
• Sidecar subtitles: SRT, VTT, or plain text for web players