agentcut 0.2.26 → 0.2.27
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +177 -177
- package/dist/studio/agents/claude.js +3 -2
- package/dist/studio/agents/claude.js.map +1 -1
- package/dist/studio/chat.js +1 -1
- package/dist/studio/chat.js.map +1 -1
- package/dist/studio/cli.js +24 -24
- package/dist/studio/download.d.ts +24 -0
- package/dist/studio/download.js +234 -0
- package/dist/studio/download.js.map +1 -0
- package/dist/studio/host.js +8 -0
- package/dist/studio/host.js.map +1 -1
- package/dist/studio/housekeeping.d.ts +2 -2
- package/dist/studio/housekeeping.js +18 -17
- package/dist/studio/housekeeping.js.map +1 -1
- package/dist/studio/release-cursor.js +21 -21
- package/dist/studio/release-video.js +10 -10
- package/dist/studio/tools.js +7 -0
- package/dist/studio/tools.js.map +1 -1
- package/package.json +56 -56
- package/template/AGENTS.md +14 -14
- package/template/components/README.md +70 -70
- package/template/components/before-after/BeforeAfter.tsx +89 -89
- package/template/components/catalog.ts +107 -107
- package/template/components/cinematic/Device3D.tsx +324 -324
- package/template/components/cinematic/devices3d/display-model.ts +184 -184
- package/template/components/cinematic/devices3d/iphone-model.ts +643 -643
- package/template/components/cinematic/devices3d/macbook-model.ts +608 -608
- package/template/components/cinematic/devices3d/models.ts +657 -657
- package/template/components/cinematic/devices3d/runtime.ts +267 -267
- package/template/components/cinematic/math.ts +100 -100
- package/template/components/color/grade-effect.ts +650 -650
- package/template/components/color/grade.ts +512 -512
- package/template/components/color/looks.ts +154 -154
- package/template/components/countdown/Countdown.tsx +166 -166
- package/template/components/dot-trace-title/DotTraceTitle.tsx +181 -181
- package/template/components/dot-trace-title/meta.json +51 -51
- package/template/components/generate/GenerateReveal.tsx +381 -381
- package/template/components/hook-card/HookCard.tsx +554 -554
- package/template/components/index.ts +199 -199
- package/template/components/lib.ts +191 -191
- package/template/components/line-diagram/LineDiagram.tsx +205 -205
- package/template/components/line-diagram/meta.json +215 -215
- package/template/components/list-card/ListCard.tsx +104 -104
- package/template/components/live-clip.ts +75 -75
- package/template/components/looks.tsx +1447 -1447
- package/template/components/loops/LoopCanvas.tsx +132 -132
- package/template/components/loops/engine.ts +184 -184
- package/template/components/loops/library.ts +667 -667
- package/template/components/loops/library2.ts +777 -777
- package/template/components/motion-loop/MotionLoop.tsx +37 -37
- package/template/components/motion-loop/meta.json +311 -311
- package/template/components/props.ts +73 -73
- package/template/components/quote-card/QuoteCard.tsx +71 -71
- package/template/components/style-layer.tsx +233 -233
- package/template/components/styles.ts +113 -113
- package/template/components/types.ts +66 -66
- package/template/components/word-pop/WordPop.tsx +59 -59
- package/template/components/word-stack/WordStack.tsx +122 -122
- package/template/presets/charcoal-sketch/PROMPT.md +48 -48
- package/template/presets/charcoal-sketch/scripts/duck_music.py +34 -34
- package/template/presets/punk-cut-paper/scenes/Motif.tsx +376 -376
- package/template/presets/punk-cut-paper/scenes/Photo.tsx +73 -73
- package/template/presets/punk-cut-paper/scenes/Ransom.tsx +119 -119
- package/template/runtime/composition.tsx +1189 -1189
- package/template/runtime/helpers.tsx +230 -230
- package/template/runtime/index.tsx +22 -22
- package/template/runtime/types.ts +315 -315
- package/template/skill/SKILL.md +186 -187
- package/template/skill/references/broll.md +44 -44
- package/template/skill/references/browser-export.md +61 -61
- package/template/skill/references/cinematic.md +206 -206
- package/template/skill/references/color.md +79 -79
- package/template/skill/references/craft.md +97 -97
- package/template/skill/references/editing-styles.md +57 -57
- package/template/skill/references/feature-launch.md +34 -34
- package/template/skill/references/footage.md +72 -72
- package/template/skill/references/format.md +142 -142
- package/template/skill/references/how-to.md +146 -146
- package/template/skill/references/launch-direction.md +89 -89
- package/template/skill/references/loops.md +151 -151
- package/template/skill/references/mcp-external.md +19 -19
- package/template/skill/references/performance.md +93 -93
- package/template/skill/references/person-cutout.md +19 -19
- package/template/skill/references/presets.md +60 -60
- package/template/skill/references/product-launch.md +134 -134
- package/template/skill/references/release-capture.md +53 -53
- package/template/skill/references/sound.md +167 -166
- package/template/skill/references/style-spy.md +160 -160
- package/template/skill/references/website-howto.md +45 -45
- package/template/starter/.studio/chat.json +1 -1
- package/template/starter/layers.json +46 -46
- package/template/starter/notes.md +4 -4
- package/template/starter/project.json +13 -13
- package/template/starter/scenes/Title.tsx +38 -38
|
@@ -1,93 +1,93 @@
|
|
|
1
|
-
# Preview performance
|
|
2
|
-
|
|
3
|
-
The user watches the live preview, which draws every scene in their browser at playback speed. A scene
|
|
4
|
-
that takes too long per frame makes the preview skip frames (choppy) and can freeze the tab while you
|
|
5
|
-
capture frames. The export doesn't skip: it renders every frame at full quality. So:
|
|
6
|
-
|
|
7
|
-
**Never lower what the render shows to make the preview smoother.** No lower resolution, fewer
|
|
8
|
-
elements on screen, skipped frames, softer images or simpler motion. Every change below must render
|
|
9
|
-
the same frames as before (check with captures); when one can't, keep the look and tell the user that
|
|
10
|
-
part previews choppy but exports perfectly.
|
|
11
|
-
|
|
12
|
-
## Measure
|
|
13
|
-
|
|
14
|
-
For website how-to videos, use the stall recovery in `website-howto.md`. Keep all required
|
|
15
|
-
format, visual, playback and audio checks. Run page-heavy tools in sequence. A timed-out request
|
|
16
|
-
can continue in the page; inspect `page.activeWork` before retrying. A small recovery probe does
|
|
17
|
-
not replace the full review or reduce final render quality.
|
|
18
|
-
|
|
19
|
-
- `agentcut_preview_check` (from/to in frames) plays a range muted in the user's preview and returns
|
|
20
|
-
per visual clip: frames played and dropped, loading stalls, the longest freeze, the most DOM elements
|
|
21
|
-
the clip drew, and a verdict: `smooth`, `choppy` from 20 % dropped, `heavy` from 40 % (and at least 6
|
|
22
|
-
dropped frames); a freeze over a second or more than 0.5 s of stalls also counts as slow. It takes as
|
|
23
|
-
long as the range plays: check the clip you changed, not the whole video every time. One check at a
|
|
24
|
-
time; it stops if the user pauses or seeks.
|
|
25
|
-
- `slowClips` lists the likeliest cause first. `sharesDropsOnly` marks a clip (often text or captions)
|
|
26
|
-
that only sits on top of a slower one: fix the other one. `slowToMount` means the clip drops frames in
|
|
27
|
-
the second before it starts, while it mounts early: its setup work is heavy (see rule 4), even when it
|
|
28
|
-
plays fine afterwards.
|
|
29
|
-
- `agentcut_status` → `page.previewPerformance` has the latest check (`lastCheck`), the user's own
|
|
30
|
-
playback (`playback`) and seek timings (`scrub`). `current: false` means it was measured on an older build.
|
|
31
|
-
- A check also returns `scrub`: how long a seek takes to reach the screen (`medianMs`, `p95Ms`, `slowClips`),
|
|
32
|
-
jumping forward and back through the range. A scene can play well and still lag on a drag, because every
|
|
33
|
-
seek draws it from nothing. Above 120 ms a drag visibly lags: the same fixes below apply, and work a scene
|
|
34
|
-
repeats on every frame (not just at mount) is the usual cause.
|
|
35
|
-
- Footage needs nothing from you: the helper makes fast-seeking copies for the preview by itself, and
|
|
36
|
-
captures and exports always read the original. Only `src` on a video clip gets one, so use a video clip
|
|
37
|
-
rather than a video inside a scene when footage is slow to scrub.
|
|
38
|
-
- Run a check after building or changing a heavy scene (walls and grids, hundreds of elements, image
|
|
39
|
-
sequences, big blurs, generated images) and before you say you're done. A first run includes loading
|
|
40
|
-
(stalls); run it again to see steady playback. A machine's own jitter varies between runs: act on a
|
|
41
|
-
`choppy` only when a second run agrees.
|
|
42
|
-
|
|
43
|
-
## Make it lighter (same pixels)
|
|
44
|
-
|
|
45
|
-
1. **Image sequences that stall.** Swapping a different picture every few frames makes the browser fetch
|
|
46
|
-
and decode a full image each time. When a check shows stalls or drops on one, encode it into one video,
|
|
47
|
-
but only if it keeps the same picture and timing:
|
|
48
|
-
- Keep the images when the frames have transparency (H.264 has none, they'd render on black), or
|
|
49
|
-
small text and hard-edged colored UI (4:2:0 color softens it).
|
|
50
|
-
- The video must show each image on exactly the frames it shows now. Images held the same number of
|
|
51
|
-
frames each: `-framerate <fps / frames per image>`. Any other timing (eased, held, reversed): write
|
|
52
|
-
the images out in display order, one file per project frame, and encode at the project fps.
|
|
53
|
-
- Encode at the frames' own size (never smaller than they ever appear), visually lossless, BT.709:
|
|
54
|
-
```
|
|
55
|
-
ffmpeg -framerate <rate> -start_number <first> -i f%03d.png \
|
|
56
|
-
-vf "pad=ceil(iw/2)*2:ceil(ih/2)*2,scale=out_color_matrix=bt709:out_range=tv,format=yuv420p" \
|
|
57
|
-
-c:v libx264 -crf 12 -preset slow -colorspace bt709 -color_primaries bt709 -color_trc bt709 \
|
|
58
|
-
-color_range tv -movflags +faststart out.mp4
|
|
59
|
-
```
|
|
60
|
-
`pad` adds one pixel on the right or bottom when a side is odd (H.264 4:2:0 needs even sizes):
|
|
61
|
-
draw the video at its padded size inside a box of the frames' size with `overflow: hidden`.
|
|
62
|
-
- ffmpeg is on PATH or in `~/.agentcut/tools/ffmpeg/`; if there's none, ask the user.
|
|
63
|
-
- Play it with `<Video>` from `@remotion/media` (`trimBefore` in frames to start inside it). Many tiles
|
|
64
|
-
of the same footage: one video per distinct clip.
|
|
65
|
-
- Capture frames across the sequence before and after and compare; if colors or edges differ,
|
|
66
|
-
keep the images.
|
|
67
|
-
2. **Repeated complex elements that never change** (a wall of the same static card): draw the piece
|
|
68
|
-
once to an image and show the copies as `<Img>`. Make it with `agentcut_capture_frames` at `scale: 2`
|
|
69
|
-
with `region` set to the piece's box in composition pixels, on a frame where it's fully drawn and
|
|
70
|
-
unobstructed; save it in `assets/` and show it at its 1x size. Keep the live version for every copy
|
|
71
|
-
that animates or changes, and for anything the camera zooms past 2x, so its text stays sharp (House
|
|
72
|
-
style).
|
|
73
|
-
3. **Draw only what's visible.** An absolutely positioned element that is fully outside the frame or at
|
|
74
|
-
opacity 0 can `return null`. Not when it sits in a flex or grid layout (opacity 0 still takes its
|
|
75
|
-
space, so its siblings would move), not when it holds `<Audio>`/`<Video>` (the sound would stop), and
|
|
76
|
-
"outside" means after the camera's transforms and including its shadow and blur.
|
|
77
|
-
4. **Compute once.** Canvases, data URLs, noise and layout math that don't depend on the frame go in
|
|
78
|
-
`useMemo` (or module scope), never in every render; nothing heavy in a component that mounts per tile.
|
|
79
|
-
Anything random must stay deterministic: `random("seed")` from `remotion`, never `Math.random()`,
|
|
80
|
-
which differs on every load and between the export's render tabs (it would flicker).
|
|
81
|
-
5. **One filter on a wrapper** instead of the same `filter`/`blur` on dozens of children, only when the
|
|
82
|
-
children don't overlap or touch (a blur or drop-shadow of the group differs where they meet), then
|
|
83
|
-
compare captures. Big animated blurs are the most expensive thing to draw: keep their area as small as
|
|
84
|
-
the look allows.
|
|
85
|
-
6. **Fewer DOM nodes for the same look.** A stack of plain divs for a skeleton UI can often be one
|
|
86
|
-
element with a `linear-gradient` background or an inline SVG that draws the same lines (check
|
|
87
|
-
`browser-export.md` for what the export draws). Never remove anything that shows.
|
|
88
|
-
7. **WebGL and canvas scenes**: `usePreviewQuality()` lowers resolution and redraw rate in the preview
|
|
89
|
-
only (`browser-export.md`); captures and the export keep full quality.
|
|
90
|
-
|
|
91
|
-
After a change, capture the same frames as before (`agentcut_capture_frames`, the export's own
|
|
92
|
-
renderer) and compare: the frames must look the same. Then run `agentcut_preview_check` again and report
|
|
93
|
-
the before/after drop rates.
|
|
1
|
+
# Preview performance
|
|
2
|
+
|
|
3
|
+
The user watches the live preview, which draws every scene in their browser at playback speed. A scene
|
|
4
|
+
that takes too long per frame makes the preview skip frames (choppy) and can freeze the tab while you
|
|
5
|
+
capture frames. The export doesn't skip: it renders every frame at full quality. So:
|
|
6
|
+
|
|
7
|
+
**Never lower what the render shows to make the preview smoother.** No lower resolution, fewer
|
|
8
|
+
elements on screen, skipped frames, softer images or simpler motion. Every change below must render
|
|
9
|
+
the same frames as before (check with captures); when one can't, keep the look and tell the user that
|
|
10
|
+
part previews choppy but exports perfectly.
|
|
11
|
+
|
|
12
|
+
## Measure
|
|
13
|
+
|
|
14
|
+
For website how-to videos, use the stall recovery in `website-howto.md`. Keep all required
|
|
15
|
+
format, visual, playback and audio checks. Run page-heavy tools in sequence. A timed-out request
|
|
16
|
+
can continue in the page; inspect `page.activeWork` before retrying. A small recovery probe does
|
|
17
|
+
not replace the full review or reduce final render quality.
|
|
18
|
+
|
|
19
|
+
- `agentcut_preview_check` (from/to in frames) plays a range muted in the user's preview and returns
|
|
20
|
+
per visual clip: frames played and dropped, loading stalls, the longest freeze, the most DOM elements
|
|
21
|
+
the clip drew, and a verdict: `smooth`, `choppy` from 20 % dropped, `heavy` from 40 % (and at least 6
|
|
22
|
+
dropped frames); a freeze over a second or more than 0.5 s of stalls also counts as slow. It takes as
|
|
23
|
+
long as the range plays: check the clip you changed, not the whole video every time. One check at a
|
|
24
|
+
time; it stops if the user pauses or seeks.
|
|
25
|
+
- `slowClips` lists the likeliest cause first. `sharesDropsOnly` marks a clip (often text or captions)
|
|
26
|
+
that only sits on top of a slower one: fix the other one. `slowToMount` means the clip drops frames in
|
|
27
|
+
the second before it starts, while it mounts early: its setup work is heavy (see rule 4), even when it
|
|
28
|
+
plays fine afterwards.
|
|
29
|
+
- `agentcut_status` → `page.previewPerformance` has the latest check (`lastCheck`), the user's own
|
|
30
|
+
playback (`playback`) and seek timings (`scrub`). `current: false` means it was measured on an older build.
|
|
31
|
+
- A check also returns `scrub`: how long a seek takes to reach the screen (`medianMs`, `p95Ms`, `slowClips`),
|
|
32
|
+
jumping forward and back through the range. A scene can play well and still lag on a drag, because every
|
|
33
|
+
seek draws it from nothing. Above 120 ms a drag visibly lags: the same fixes below apply, and work a scene
|
|
34
|
+
repeats on every frame (not just at mount) is the usual cause.
|
|
35
|
+
- Footage needs nothing from you: the helper makes fast-seeking copies for the preview by itself, and
|
|
36
|
+
captures and exports always read the original. Only `src` on a video clip gets one, so use a video clip
|
|
37
|
+
rather than a video inside a scene when footage is slow to scrub.
|
|
38
|
+
- Run a check after building or changing a heavy scene (walls and grids, hundreds of elements, image
|
|
39
|
+
sequences, big blurs, generated images) and before you say you're done. A first run includes loading
|
|
40
|
+
(stalls); run it again to see steady playback. A machine's own jitter varies between runs: act on a
|
|
41
|
+
`choppy` only when a second run agrees.
|
|
42
|
+
|
|
43
|
+
## Make it lighter (same pixels)
|
|
44
|
+
|
|
45
|
+
1. **Image sequences that stall.** Swapping a different picture every few frames makes the browser fetch
|
|
46
|
+
and decode a full image each time. When a check shows stalls or drops on one, encode it into one video,
|
|
47
|
+
but only if it keeps the same picture and timing:
|
|
48
|
+
- Keep the images when the frames have transparency (H.264 has none, they'd render on black), or
|
|
49
|
+
small text and hard-edged colored UI (4:2:0 color softens it).
|
|
50
|
+
- The video must show each image on exactly the frames it shows now. Images held the same number of
|
|
51
|
+
frames each: `-framerate <fps / frames per image>`. Any other timing (eased, held, reversed): write
|
|
52
|
+
the images out in display order, one file per project frame, and encode at the project fps.
|
|
53
|
+
- Encode at the frames' own size (never smaller than they ever appear), visually lossless, BT.709:
|
|
54
|
+
```
|
|
55
|
+
ffmpeg -framerate <rate> -start_number <first> -i f%03d.png \
|
|
56
|
+
-vf "pad=ceil(iw/2)*2:ceil(ih/2)*2,scale=out_color_matrix=bt709:out_range=tv,format=yuv420p" \
|
|
57
|
+
-c:v libx264 -crf 12 -preset slow -colorspace bt709 -color_primaries bt709 -color_trc bt709 \
|
|
58
|
+
-color_range tv -movflags +faststart out.mp4
|
|
59
|
+
```
|
|
60
|
+
`pad` adds one pixel on the right or bottom when a side is odd (H.264 4:2:0 needs even sizes):
|
|
61
|
+
draw the video at its padded size inside a box of the frames' size with `overflow: hidden`.
|
|
62
|
+
- ffmpeg is on PATH or in `~/.agentcut/tools/ffmpeg/`; if there's none, ask the user.
|
|
63
|
+
- Play it with `<Video>` from `@remotion/media` (`trimBefore` in frames to start inside it). Many tiles
|
|
64
|
+
of the same footage: one video per distinct clip.
|
|
65
|
+
- Capture frames across the sequence before and after and compare; if colors or edges differ,
|
|
66
|
+
keep the images.
|
|
67
|
+
2. **Repeated complex elements that never change** (a wall of the same static card): draw the piece
|
|
68
|
+
once to an image and show the copies as `<Img>`. Make it with `agentcut_capture_frames` at `scale: 2`
|
|
69
|
+
with `region` set to the piece's box in composition pixels, on a frame where it's fully drawn and
|
|
70
|
+
unobstructed; save it in `assets/` and show it at its 1x size. Keep the live version for every copy
|
|
71
|
+
that animates or changes, and for anything the camera zooms past 2x, so its text stays sharp (House
|
|
72
|
+
style).
|
|
73
|
+
3. **Draw only what's visible.** An absolutely positioned element that is fully outside the frame or at
|
|
74
|
+
opacity 0 can `return null`. Not when it sits in a flex or grid layout (opacity 0 still takes its
|
|
75
|
+
space, so its siblings would move), not when it holds `<Audio>`/`<Video>` (the sound would stop), and
|
|
76
|
+
"outside" means after the camera's transforms and including its shadow and blur.
|
|
77
|
+
4. **Compute once.** Canvases, data URLs, noise and layout math that don't depend on the frame go in
|
|
78
|
+
`useMemo` (or module scope), never in every render; nothing heavy in a component that mounts per tile.
|
|
79
|
+
Anything random must stay deterministic: `random("seed")` from `remotion`, never `Math.random()`,
|
|
80
|
+
which differs on every load and between the export's render tabs (it would flicker).
|
|
81
|
+
5. **One filter on a wrapper** instead of the same `filter`/`blur` on dozens of children, only when the
|
|
82
|
+
children don't overlap or touch (a blur or drop-shadow of the group differs where they meet), then
|
|
83
|
+
compare captures. Big animated blurs are the most expensive thing to draw: keep their area as small as
|
|
84
|
+
the look allows.
|
|
85
|
+
6. **Fewer DOM nodes for the same look.** A stack of plain divs for a skeleton UI can often be one
|
|
86
|
+
element with a `linear-gradient` background or an inline SVG that draws the same lines (check
|
|
87
|
+
`browser-export.md` for what the export draws). Never remove anything that shows.
|
|
88
|
+
7. **WebGL and canvas scenes**: `usePreviewQuality()` lowers resolution and redraw rate in the preview
|
|
89
|
+
only (`browser-export.md`); captures and the export keep full quality.
|
|
90
|
+
|
|
91
|
+
After a change, capture the same frames as before (`agentcut_capture_frames`, the export's own
|
|
92
|
+
renderer) and compare: the frames must look the same. Then run `agentcut_preview_check` again and report
|
|
93
|
+
the before/after drop rates.
|
|
@@ -1,19 +1,19 @@
|
|
|
1
|
-
# Graphics behind a person
|
|
2
|
-
|
|
3
|
-
Use agentcut_cutout_person when the user wants text or a graphic behind someone, or the person in front of something. Do not install segmentation packages, write background-removal scripts, export frames or use image APIs.
|
|
4
|
-
|
|
5
|
-
The stack, top to bottom: the person cutout, the graphic, the original footage. The footage stays in place, so a slightly loose edge only shows where the graphic touches the person.
|
|
6
|
-
|
|
7
|
-
1. Add the graphic on a track above the footage first, and note its clip ID and frames.
|
|
8
|
-
2. Call agentcut_cutout_person with the footage clipId, fromFrame and durationInFrames covering only the frames where the graphic is on screen, and behind with the graphic's clip ID. Cutting out only those frames is much faster than the whole clip. The first run also downloads a model of about 26 MB.
|
|
9
|
-
3. Check captures at the start, middle and end of the range, and where the graphic overlaps the person. Look at hair, hands and fast movement.
|
|
10
|
-
|
|
11
|
-
The tool adds a muted video clip to a track named Person cutout. Keep captions on a track above the cutout so the person doesn't cover them.
|
|
12
|
-
|
|
13
|
-
- To grade a person apart from their background, use `agentcut_grade_subject` instead (see `references/color.md`).
|
|
14
|
-
- It copies the footage clip's trim, speed, box, crop, fit and grade once. After changing those, or moving the footage, run the tool again with the same range; it replaces the earlier cutout on that range.
|
|
15
|
-
- A range can't start or end inside an enter or exit animation of the footage, and a project style camera needs the whole clip: the tool returns needs_full_clip, so omit the range.
|
|
16
|
-
- Works on the person in front; it can't choose between people, and it doesn't handle objects. Fast movement can leave soft or smeared edges. no_person means most of the range has nobody clearly visible.
|
|
17
|
-
- Without WebGPU the tool is several times slower and accepts up to 1,800 frames at a time; with WebGPU, 5,400.
|
|
18
|
-
- Frames stay on this computer. Scene-embedded video and remote URLs aren't supported.
|
|
19
|
-
- Files go to assets/cutouts/. Use Undo, or delete the clip, to remove a cutout.
|
|
1
|
+
# Graphics behind a person
|
|
2
|
+
|
|
3
|
+
Use agentcut_cutout_person when the user wants text or a graphic behind someone, or the person in front of something. Do not install segmentation packages, write background-removal scripts, export frames or use image APIs.
|
|
4
|
+
|
|
5
|
+
The stack, top to bottom: the person cutout, the graphic, the original footage. The footage stays in place, so a slightly loose edge only shows where the graphic touches the person.
|
|
6
|
+
|
|
7
|
+
1. Add the graphic on a track above the footage first, and note its clip ID and frames.
|
|
8
|
+
2. Call agentcut_cutout_person with the footage clipId, fromFrame and durationInFrames covering only the frames where the graphic is on screen, and behind with the graphic's clip ID. Cutting out only those frames is much faster than the whole clip. The first run also downloads a model of about 26 MB.
|
|
9
|
+
3. Check captures at the start, middle and end of the range, and where the graphic overlaps the person. Look at hair, hands and fast movement.
|
|
10
|
+
|
|
11
|
+
The tool adds a muted video clip to a track named Person cutout. Keep captions on a track above the cutout so the person doesn't cover them.
|
|
12
|
+
|
|
13
|
+
- To grade a person apart from their background, use `agentcut_grade_subject` instead (see `references/color.md`).
|
|
14
|
+
- It copies the footage clip's trim, speed, box, crop, fit and grade once. After changing those, or moving the footage, run the tool again with the same range; it replaces the earlier cutout on that range.
|
|
15
|
+
- A range can't start or end inside an enter or exit animation of the footage, and a project style camera needs the whole clip: the tool returns needs_full_clip, so omit the range.
|
|
16
|
+
- Works on the person in front; it can't choose between people, and it doesn't handle objects. Fast movement can leave soft or smeared edges. no_person means most of the range has nobody clearly visible.
|
|
17
|
+
- Without WebGPU the tool is several times slower and accepts up to 1,800 frames at a time; with WebGPU, 5,400.
|
|
18
|
+
- Frames stay on this computer. Scene-embedded video and remote URLs aren't supported.
|
|
19
|
+
- Files go to assets/cutouts/. Use Undo, or delete the clip, to remove a cutout.
|
|
@@ -1,60 +1,60 @@
|
|
|
1
|
-
# Video style presets
|
|
2
|
-
|
|
3
|
-
These are scene/asset kits. For Save current style, saved editing rules, or applying an editing style to different
|
|
4
|
-
footage, read `references/editing-styles.md` instead. Existing kits stay in `library/presets/`.
|
|
5
|
-
|
|
6
|
-
A video style is a recipe for a whole video that isn't a talking head: a music or lyric video, an explainer, a
|
|
7
|
-
promo look. It lives in the workspace library, one folder per preset:
|
|
8
|
-
|
|
9
|
-
```
|
|
10
|
-
library/presets/<id>/
|
|
11
|
-
PROMPT.md the recipe: what to copy, the look, the workflow. Blanks in [brackets].
|
|
12
|
-
preset.json name, description, kind, preview/poster file names, and the blanks as form fields
|
|
13
|
-
preview.mp4 ~10 s of a video made with it (16:9, 960×540, no audio needed)
|
|
14
|
-
poster.jpg one strong frame from it
|
|
15
|
-
scenes/ the scene files the recipe copies into a project
|
|
16
|
-
assets/ textures, fonts and other files the scenes need
|
|
17
|
-
*.example.* optional: a script or layers.json that seeds the first timeline
|
|
18
|
-
```
|
|
19
|
-
|
|
20
|
-
The style gallery (Styles tab → Video styles) lists every folder with a PROMPT.md, plays its preview, and shows a
|
|
21
|
-
form built from `preset.json` fields. When the user fills it in, you get a message like:
|
|
22
|
-
|
|
23
|
-
> Make a video with the "Punk Cut-Paper" video style: read library/presets/punk-cut-paper/PROMPT.md and follow it,
|
|
24
|
-
> with these values for its blanks: - Song: assets/song.mp3 - Photos: … - Lyrics: transcribe them. Make it as a new
|
|
25
|
-
> project.
|
|
26
|
-
|
|
27
|
-
## Using a preset
|
|
28
|
-
|
|
29
|
-
1. Read the preset's PROMPT.md and fill its blanks with the values given (ask only for what's missing and needed).
|
|
30
|
-
2. "As a new project": create it with `agentcut_create_project`, then copy the preset's `scenes/` and
|
|
31
|
-
`assets/` into it (and any song/photos the user picked from another project or the library). "In this project":
|
|
32
|
-
copy them into the open project and replace its timeline.
|
|
33
|
-
3. Follow the recipe's look and workflow exactly; it was written from a finished video, so don't improvise a new look.
|
|
34
|
-
|
|
35
|
-
## Saving a video style
|
|
36
|
-
|
|
37
|
-
Use this workflow only when the user explicitly asks to save a scene/asset kit. The gallery's Save current style
|
|
38
|
-
action saves editing instructions instead; follow `references/editing-styles.md` for it.
|
|
39
|
-
|
|
40
|
-
1. Copy the project's scene files and the assets the look needs (textures, fonts, overlays: not the user's song or
|
|
41
|
-
personal photos) into `library/presets/<id>/`. Make the scenes generic: no names, personal facts or lyrics baked in
|
|
42
|
-
(put example content in an `*.example.*` file).
|
|
43
|
-
2. Write PROMPT.md: what to copy, fonts for project.json, every rule of the look (layout, motion, timing to beats or
|
|
44
|
-
words, colors, overlays), and the workflow. Put what changes per video in [brackets] (song, photos, title, about,
|
|
45
|
-
lyrics, …).
|
|
46
|
-
3. Write preset.json:
|
|
47
|
-
```json
|
|
48
|
-
{ "name": "Punk Cut-Paper", "description": "One sentence on the look.", "kind": "Music video",
|
|
49
|
-
"preview": "preview.mp4", "poster": "poster.jpg",
|
|
50
|
-
"fields": [
|
|
51
|
-
{ "id": "song", "label": "Song", "type": "audio", "required": true },
|
|
52
|
-
{ "id": "photos", "label": "Photos", "type": "images" },
|
|
53
|
-
{ "id": "title", "label": "Title", "type": "text", "placeholder": "HAPPY BIRTHDAY SAM" },
|
|
54
|
-
{ "id": "lyrics", "label": "Lyrics", "type": "longtext", "placeholder": "Paste the lyrics, or leave empty" }
|
|
55
|
-
] }
|
|
56
|
-
```
|
|
57
|
-
Field types: `text`, `longtext`, `audio` (the user picks an audio file), `images` (the user picks photos).
|
|
58
|
-
4. preview.mp4 and poster.jpg: cut ~10 s from the latest render in `renders/` (960×540, H.264, no audio) and grab one
|
|
59
|
-
frame, choosing a stretch without personal photos when you can. With no render yet, export one first or ask.
|
|
60
|
-
5. Tell the user it's in the Styles tab under Video styles.
|
|
1
|
+
# Video style presets
|
|
2
|
+
|
|
3
|
+
These are scene/asset kits. For Save current style, saved editing rules, or applying an editing style to different
|
|
4
|
+
footage, read `references/editing-styles.md` instead. Existing kits stay in `library/presets/`.
|
|
5
|
+
|
|
6
|
+
A video style is a recipe for a whole video that isn't a talking head: a music or lyric video, an explainer, a
|
|
7
|
+
promo look. It lives in the workspace library, one folder per preset:
|
|
8
|
+
|
|
9
|
+
```
|
|
10
|
+
library/presets/<id>/
|
|
11
|
+
PROMPT.md the recipe: what to copy, the look, the workflow. Blanks in [brackets].
|
|
12
|
+
preset.json name, description, kind, preview/poster file names, and the blanks as form fields
|
|
13
|
+
preview.mp4 ~10 s of a video made with it (16:9, 960×540, no audio needed)
|
|
14
|
+
poster.jpg one strong frame from it
|
|
15
|
+
scenes/ the scene files the recipe copies into a project
|
|
16
|
+
assets/ textures, fonts and other files the scenes need
|
|
17
|
+
*.example.* optional: a script or layers.json that seeds the first timeline
|
|
18
|
+
```
|
|
19
|
+
|
|
20
|
+
The style gallery (Styles tab → Video styles) lists every folder with a PROMPT.md, plays its preview, and shows a
|
|
21
|
+
form built from `preset.json` fields. When the user fills it in, you get a message like:
|
|
22
|
+
|
|
23
|
+
> Make a video with the "Punk Cut-Paper" video style: read library/presets/punk-cut-paper/PROMPT.md and follow it,
|
|
24
|
+
> with these values for its blanks: - Song: assets/song.mp3 - Photos: … - Lyrics: transcribe them. Make it as a new
|
|
25
|
+
> project.
|
|
26
|
+
|
|
27
|
+
## Using a preset
|
|
28
|
+
|
|
29
|
+
1. Read the preset's PROMPT.md and fill its blanks with the values given (ask only for what's missing and needed).
|
|
30
|
+
2. "As a new project": create it with `agentcut_create_project`, then copy the preset's `scenes/` and
|
|
31
|
+
`assets/` into it (and any song/photos the user picked from another project or the library). "In this project":
|
|
32
|
+
copy them into the open project and replace its timeline.
|
|
33
|
+
3. Follow the recipe's look and workflow exactly; it was written from a finished video, so don't improvise a new look.
|
|
34
|
+
|
|
35
|
+
## Saving a video style
|
|
36
|
+
|
|
37
|
+
Use this workflow only when the user explicitly asks to save a scene/asset kit. The gallery's Save current style
|
|
38
|
+
action saves editing instructions instead; follow `references/editing-styles.md` for it.
|
|
39
|
+
|
|
40
|
+
1. Copy the project's scene files and the assets the look needs (textures, fonts, overlays: not the user's song or
|
|
41
|
+
personal photos) into `library/presets/<id>/`. Make the scenes generic: no names, personal facts or lyrics baked in
|
|
42
|
+
(put example content in an `*.example.*` file).
|
|
43
|
+
2. Write PROMPT.md: what to copy, fonts for project.json, every rule of the look (layout, motion, timing to beats or
|
|
44
|
+
words, colors, overlays), and the workflow. Put what changes per video in [brackets] (song, photos, title, about,
|
|
45
|
+
lyrics, …).
|
|
46
|
+
3. Write preset.json:
|
|
47
|
+
```json
|
|
48
|
+
{ "name": "Punk Cut-Paper", "description": "One sentence on the look.", "kind": "Music video",
|
|
49
|
+
"preview": "preview.mp4", "poster": "poster.jpg",
|
|
50
|
+
"fields": [
|
|
51
|
+
{ "id": "song", "label": "Song", "type": "audio", "required": true },
|
|
52
|
+
{ "id": "photos", "label": "Photos", "type": "images" },
|
|
53
|
+
{ "id": "title", "label": "Title", "type": "text", "placeholder": "HAPPY BIRTHDAY SAM" },
|
|
54
|
+
{ "id": "lyrics", "label": "Lyrics", "type": "longtext", "placeholder": "Paste the lyrics, or leave empty" }
|
|
55
|
+
] }
|
|
56
|
+
```
|
|
57
|
+
Field types: `text`, `longtext`, `audio` (the user picks an audio file), `images` (the user picks photos).
|
|
58
|
+
4. preview.mp4 and poster.jpg: cut ~10 s from the latest render in `renders/` (960×540, H.264, no audio) and grab one
|
|
59
|
+
frame, choosing a stretch without personal photos when you can. With no render yet, export one first or ask.
|
|
60
|
+
5. Tell the user it's in the Styles tab under Video styles.
|