Can I change the words and colors?

title is a JSON input and palettes/semantic overrides remain available. The current title itself is fixed white and fitted to 1730 px independently of the UI palette; changing that treatment requires a scene edit.

Do I need to supply footage?

Yes. Bring a continuous 6.8-second source take and an aligned transparent cutout, or allow automatic local cutting. Both base and foreground are video layers; a static portrait does not replace the required matte here.

Can I revise the motion?

Yes. Preserve the title between exactly matched video layers, its brief punch/settle and later clearance. Re-render after any crop or scale changes and inspect registration in motion.

What if automatic cutting fails or leaves rough edges?

Review or run the pinned remove-background command separately, then supply a better aligned matte. The generator caches successful output by source checksum, but that does not guarantee visual edge quality.

The guide recommends 4K footage and --resolution landscape-4k on the render command. Inspect the actual render and cutout edges at that resolution; a 1080p preview does not verify 4K quality.

CreatorMotion · Title Behind Subject

Previews from the publisher

Give a short title scale while keeping the real presenter in front of the letters. A base video, dark tint, oversized title and aligned transparent subject layer create the depth effect; the title punches in, settles and later clears. This 6.8-second shot requires matched source/cutout material rather than simulating a silhouette with an unrelated image.

What you get

A local generator, editable HyperFrames layered shot and motion/review instructions for a 1920×1080, 30 fps segment. Generate and render the title treatment over your supplied continuous presenter footage.
One editable title plus --media and a required aligned presenter-matte alpha video. The base and foreground share the same crop, scale and positioning; the title sits between them. A static portrait is not a supported substitute in this scene.
A title punch from 0.85 to 1.12 seconds, small settle through 1.30 seconds and rapid clearance from 5.55 seconds. White slanted SVG text uses a painted width of 1730 px independently of the UI palette, while the presenter continues uninterrupted.
Four styles keep the same data and choreography: White & Yellow (white-yellow), Charcoal & Lime (charcoal-lime), Paper & Orange (paper-orange), and Black & White (midnight-ice). Semantic style overrides are supported; check contrast after changing them. Dark canvases use a fine grid. Local fonts include Inter, IBM Plex Mono for measurements and ordinals, Smiley Sans and Noto Sans SC. Chinese text selects Smiley Sans automatically, including accompanying Latin text; --zh-font noto-sans-sc selects Noto Sans SC, while --font takes precedence for a supplied font.

Before you choose

Plan the phrase and source framing together so foreground occlusion preserves both the presenter and the meaning of the title. Review matched edges before judging the typography; the depth effect depends on accurate layers, not merely large letters.

Good fit

A creator introducing a short emphatic phrase with genuine foreground occlusion around a presenter.
A teaching or chapter beat where the person should remain visible while the title occupies more of the frame.
A series with repeatable base/matte alignment, large type and a quiet speaker-only ending.

Not for

Creating a missing presenter performance, using an arbitrary still silhouette or assuming unmatched video layers will align.
Long prose, automatically protecting every letter from occlusion or guaranteed clean background removal for any footage.
Automatic full-film editing, generated soundtrack or copying another creator’s identity and source materials.

How to use

You say

/creatormotion-type-behind-subject Turn my supplied explanation into a white-yellow B-roll scene using the Title Behind Subject effect. Use the text, data and media I provide, and ask for any required input that is missing.

Write one short title. Keep it concise: the SVG fixes the painted text width rather than preserving natural letter proportions for any length. Review actual pixels and intentional foreground occlusion, not merely the inspector’s uncompressed glyph measurement.
Supply a continuous local --media take covering at least 6.8 seconds and its aligned cutout via --matte FILE or presenter-matte in --images. Source and matte must share frames, duration, framing and the same crop. Alpha-bearing footage can be recognized as its own cutout; a separate portrait fallback is not allowed by this required-matte policy.
If no cutout exists, leave --matte auto so the generator invokes pinned HyperFrames remove-background locally and caches by source checksum. The guide describes a first-use model download. --matte off only disables automatic cutting; it does not remove the effect’s need for a matte. Inspect hair, hands and fast motion and supply a better cutout when needed.
From the extracted package, prepare Node.js 22+, FFmpeg and HyperFrames CLI 0.8.30. Run node scripts/create.mjs --out PROJECT --data CONTENT.json --style white-yellow, substituting another documented palette if wanted. Read the motion specification and image policy before adapting the scene; use semantic --overrides for deliberate style changes. Add --media /absolute/PRESENTER.mp4 and, when available, --matte /absolute/MATTE.webm. Keep the source continuous and muted; do not loop a shorter take.
Inspect the editable project at entry, each motion handoff and final hold, then render with the command printed by the generator. Check the resulting frames for text collisions, correct numbers and readable contrast. Keep at least 1.2 seconds for the final reading when editing, and compare all four palettes with identical content after structural changes. Narration, music and assembly into a longer film remain separate work. Inspect base/matte registration at every title edge, the 140% matched crop and the final speaker view. Provenance/checksums record cutout resolution, not segmentation quality. Do not claim narration synchronization from a silent preview.

Requirements and costs

You bring: title: provide your own content or retain the illustrative default.
Dependencies: HyperFrames CLI 0.8.30 or compatible — required · Node.js 22 or newer — required · FFmpeg — required
Outside costs: Not collected yet
Tools used: HyperFrames
Publisher testing: v1.0.2 — Claude Code / Claude Desktop / Cowork / Codex / ChatGPT