Open-source AI podcast clipper. Generate vertical clips with face tracking and burned-in captions. CLI, MCP server, web app.
Python ai-toolsauto-captionscaptionscliffmpegmcpmcp-serveropusclip-alternativepodcast-clipspodcastsshortform-videovertical-videowhisper
Draw the topic chip, progress bar and cards the templates already offer (#189)
Nika Siradzed8b596f8a84c
diff --git a/backend/cli.py b/backend/cli.pyindex 4b9af34..37d1dc8 100644--- a/backend/cli.py+++ b/backend/cli.py@@ -475,6 +475,21 @@ def cmd_studio(args):"--intro-seconds", str(args.intro_seconds),"--outro-seconds", str(args.outro_seconds),]+ # Only when asked for, so a studio cut that names none of these is the+ # same command it was before they existed.+ if getattr(args, "topic", None):+ cmd += ["--topic", args.topic,+ "--topic-position", getattr(args, "topic_position", "top-left") or "top-left"]+ if getattr(args, "progress", False):+ cmd += ["--progress"]+ if getattr(args, "progress_color", None):+ cmd += ["--progress-color", args.progress_color]+ if getattr(args, "cards", None):+ cmd += ["--cards", args.cards]+ if getattr(args, "brand", None):+ cmd += ["--brand", args.brand]+ if getattr(args, "font_family", None):+ cmd += ["--font-family", args.font_family]if args.outro_title is not None:cmd += ["--outro-title", args.outro_title]if args.platforms is not None:@@ -676,6 +691,26 @@ def cmd_process(args):config["logo_position"] = args.logo_positionif getattr(args, "logo_scale", None) is not None:config["logo_scale"] = args.logo_scale+ # Switched-on parts of the look. Absent stays absent: the renderer draws+ # nothing for a key that is not here, which is what every render did before+ # these flags existed.+ if getattr(args, "topic", None):+ config["topic"] = {+ "label": args.topic,+ "position": getattr(args, "topic_position", "top-left") or "top-left",+ }+ if getattr(args, "progress", False):+ config["progress"] = (+ {"color": args.progress_color} if getattr(args, "progress_color", None) else {}+ )+ if getattr(args, "brand", None):+ try:+ config["brand"] = json.loads(args.brand)+ except (ValueError, TypeError):+ print(" Warning: --brand is not valid JSON; using the default colours",+ file=sys.stderr)+ if getattr(args, "font_family", None):+ config["font_family"] = args.font_familyif getattr(args, "no_outro", False):config["outro_path"] = ""elif args.outro:@@ -1178,6 +1213,10 @@ def cmd_process(args):caption_font_scale=config.get("caption_font_scale", 100),logo_position=config.get("logo_position", "top-left"),logo_scale=config.get("logo_scale", 1.0),+ topic=config.get("topic"),+ progress=config.get("progress"),+ brand=config.get("brand"),+ font_family=config.get("font_family"),crop_strategy=config.get("crop_strategy", "face"),format=config.get("format", "vertical"),transcript_words=words,@@ -1403,6 +1442,10 @@ def cmd_process(args):caption_font_scale=config.get("caption_font_scale", 100),logo_position=config.get("logo_position", "top-left"),logo_scale=config.get("logo_scale", 1.0),+ topic=config.get("topic"),+ progress=config.get("progress"),+ brand=config.get("brand"),+ font_family=config.get("font_family"),crop_strategy=config.get("crop_strategy", "face"),format=config.get("format", "vertical"),transcript_words=words,@@ -4251,6 +4294,19 @@ def main():help="Logo placement (default: top-left)")proc.add_argument("--logo-scale", type=float, choices=[0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, 1.0, 1.05, 1.1, 1.15, 1.2, 1.25, 1.3, 1.35, 1.4, 1.45, 1.5],help="Logo size multiplier (default: 1)")+ # The parts a show's look switches on. Each is off unless asked for, so a+ # command that names none of them cuts exactly what it cut before.+ proc.add_argument("--topic", help="Standing label saying what the clip is about")+ proc.add_argument("--topic-position", dest="topic_position", choices=["top-left", "top-center", "top-right", "bottom-left", "bottom-center", "bottom-right"],+ default="top-left", help="Where the topic label sits (default: top-left)")+ proc.add_argument("--progress", action="store_true",+ help="Draw how much of the clip is left along the bottom edge")+ proc.add_argument("--progress-color", dest="progress_color",+ help="Colour of the progress bar, e.g. '#3B9CFF'")+ proc.add_argument("--brand", help="Show colours as JSON: "+ '{"accent":"#4C9DF5","ink":"#FFFFFF","surface":"#0A0D14"}')+ proc.add_argument("--font-family", dest="font_family",+ help="The show's own typeface, ahead of the built-in stack")proc.add_argument("--outro", help="Outro video (asset name or path)")proc.add_argument("--name-card", dest="name_card",help="Lower third naming the speaker, shown for the first seconds")@@ -4323,6 +4379,17 @@ def main():studio.add_argument("--logo", help="Logo image (asset name or path)")studio.add_argument("--logo-position", choices=["top-left", "top-center", "top-right", "bottom-left", "bottom-center", "bottom-right"], default="top-left")studio.add_argument("--logo-scale", type=float, default=1.0)+ studio.add_argument("--topic", help="Standing label saying what the clip is about")+ studio.add_argument("--topic-position", dest="topic_position", choices=["top-left", "top-center", "top-right", "bottom-left", "bottom-center", "bottom-right"],+ default="top-left")+ studio.add_argument("--progress", action="store_true",+ help="Draw how much of the clip is left along the bottom edge")+ studio.add_argument("--progress-color", dest="progress_color")+ studio.add_argument("--cards", help="On-screen cards as JSON, each with kind/start/end")+ studio.add_argument("--brand", help="Show colours as JSON: "+ '{"accent":"#4C9DF5","ink":"#FFFFFF","surface":"#0A0D14"}')+ studio.add_argument("--font-family", dest="font_family",+ help="The show's own typeface, ahead of the built-in stack")studio.add_argument("--intro", help="Intro video (asset name or path)")studio.add_argument("--outro", help="Outro video (asset name or path)")studio.add_argument("--name-card", dest="name_card",diff --git a/backend/clip_studio.py b/backend/clip_studio.pyindex 0e19e75..6056b2f 100644--- a/backend/clip_studio.py+++ b/backend/clip_studio.py@@ -119,9 +119,21 @@ def _find_paragraph(words: list, phrase: str) -> tuple[float, float]:return start, end+def _json_arg(raw, name):+ """A JSON argument, or nothing. Malformed loses the part, never the clip."""+ if not raw:+ return None+ try:+ return json.loads(raw)+ except (ValueError, TypeError):+ print(f" Warning: {name} is not valid JSON; ignoring it", file=sys.stderr, flush=True)+ return None++def _render_fragment(video, start, end, words, style, crop, title, out_dir, fmt="vertical",logo=None, name_card=None, motion=None, caption_position="auto",- caption_scale=1.0, logo_position="top-left", logo_scale=1.0):+ caption_scale=1.0, logo_position="top-left", logo_scale=1.0,+ topic=None, progress=None, cards=None, brand=None, font_family=None):"""Render the fragment with face-crop + captions via the existing engine."""from services.clip_generator import generate_clipprint(f" [fragment] rendering {start:.1f}s–{end:.1f}s ({style}, crop={crop}, {fmt})", flush=True)@@ -133,6 +145,7 @@ def _render_fragment(video, start, end, words, style, crop, title, out_dir, fmt=crop_strategy=crop, format=fmt,transcript_words=words, title=title, output_dir=out_dir,logo_path=logo, name_card=name_card, motion=motion,+ topic=topic, progress=progress, cards=cards, brand=brand, font_family=font_family,clean_fillers=True, allow_ass_fallback=True,progress_callback=lambda p, m: print(f" {p}% {m}", flush=True),)@@ -229,6 +242,20 @@ def main():ap.add_argument("--logo", default=None, help="Logo image (asset name or path)")ap.add_argument("--logo-position", default="top-left", choices=["top-left", "top-center", "top-right", "bottom-left", "bottom-center", "bottom-right"])ap.add_argument("--logo-scale", type=float, default=1.0)+ ap.add_argument("--topic", default=None,+ help="Standing label saying what the clip is about")+ ap.add_argument("--topic-position", default="top-left",+ choices=["top-left", "top-center", "top-right",+ "bottom-left", "bottom-center", "bottom-right"])+ ap.add_argument("--progress", action="store_true",+ help="Draw how much of the clip is left along the bottom edge")+ ap.add_argument("--progress-color", default=None)+ ap.add_argument("--cards", default=None,+ help="On-screen cards as JSON, each with kind/start/end")+ ap.add_argument("--brand", default=None,+ help='Show colours as JSON: {"accent":"#4C9DF5","ink":"#FFF","surface":"#000"}')+ ap.add_argument("--font-family", default=None,+ help="The show's own typeface, ahead of the built-in stack")ap.add_argument("--intro", default=None, help="Intro video (asset name or path)")ap.add_argument("--outro", default=None, help="Outro video (asset name or path)")# bookends (defaults are None so we can tell what the user explicitly set;@@ -327,6 +354,14 @@ def main():caption_scale=max(0.7, min(1.5, args.caption_scale)),logo_position=args.logo_position,logo_scale=max(0.5, min(1.5, args.logo_scale)),+ topic=({"label": args.topic, "position": args.topic_position} if args.topic else None),+ progress=(+ ({"color": args.progress_color} if args.progress_color else {})+ if args.progress else None+ ),+ cards=_json_arg(args.cards, "--cards"),+ brand=_json_arg(args.brand, "--brand"),+ font_family=args.font_family,)platforms = [p.strip() for p in platforms_str.split(",") if p.strip()]diff --git a/backend/main.py b/backend/main.pyindex 52ec787..da7e237 100644--- a/backend/main.py+++ b/backend/main.py@@ -172,6 +172,11 @@ def handle_create_clip(task_id: str, params: dict):caption_position=params.get("caption_position", "auto"),caption_font_scale=params.get("caption_font_scale", 100),logo_position=params.get("logo_position", "top-left"),+ topic=params.get("topic"),+ progress=params.get("progress"),+ cards=params.get("cards"),+ brand=params.get("brand"),+ font_family=params.get("font_family"),crop_strategy=params.get("crop_strategy", "face"),format=params.get("format", "vertical"),crop_keyframes=params.get("crop_keyframes"),@@ -234,6 +239,11 @@ def handle_batch_clips(task_id: str, params: dict):caption_position=clip.get("caption_position", params.get("caption_position", "auto")),caption_font_scale=clip.get("caption_font_scale", params.get("caption_font_scale", 100)),logo_position=clip.get("logo_position", params.get("logo_position", "top-left")),+ topic=clip.get("topic", params.get("topic")),+ progress=clip.get("progress", params.get("progress")),+ cards=clip.get("cards", params.get("cards")),+ brand=clip.get("brand", params.get("brand")),+ font_family=clip.get("font_family", params.get("font_family")),crop_strategy=clip.get("crop_strategy", "face"),format=clip.get("format", params.get("format", "vertical")),transcript_words=params.get("transcript_words", []),diff --git a/backend/services/clip_generator.py b/backend/services/clip_generator.pyindex ab4c078..a9a0b87 100644--- a/backend/services/clip_generator.py+++ b/backend/services/clip_generator.py@@ -496,6 +496,15 @@ def _report_remotion_failure(print(f" Remotion {stage} {label}: {line[:200]}", file=sys.stderr, flush=True)+# Watermark.tsx's own numbers, so the ffmpeg fallback puts the mark where the+# Remotion renderer would have put it. Authored against a 1920-tall frame and+# scaled by the height a render actually produced, exactly as captionScale does.+REMOTION_REFERENCE_HEIGHT = 1920+REMOTION_LOGO_HEIGHT = 126+REMOTION_LOGO_INSET = 180+REMOTION_LOGO_EDGE = 108++def _kept_caption_overlay_path(output_path: str) -> str:base, _ = os.path.splitext(os.path.abspath(output_path))return f"{base}_captions.mov"@@ -515,6 +524,11 @@ def _render_with_remotion(caption_font_scale: int = 100,logo_position: str = "top-left",logo_scale: float = 1.0,+ topic: Optional[dict] = None,+ progress: Optional[dict] = None,+ cards: Optional[list] = None,+ brand: Optional[dict] = None,+ font_family: Optional[str] = None,) -> tuple[bool, Optional[str]]:"""Render captions using Remotion. Returns (success, optional_prores_overlay_path).@@ -656,6 +670,26 @@ def _render_with_remotion(cmd.extend(["--name-card-accent", str(name_card["accent"])])if motion:cmd.extend(["--motion", json.dumps(motion)])+ # The parts a template switches on. Each is only sent when it has a+ # value, so a caller that passes none renders what it always rendered.+ if topic and topic.get("label"):+ cmd.extend(["--topic", str(topic["label"])])+ if topic.get("position"):+ cmd.extend(["--topic-position", str(topic["position"])])+ if topic.get("color"):+ cmd.extend(["--topic-color", str(topic["color"])])+ if topic.get("background"):+ cmd.extend(["--topic-background", str(topic["background"])])+ if progress:+ cmd.append("--progress")+ if progress.get("color"):+ cmd.extend(["--progress-color", str(progress["color"])])+ if cards:+ cmd.extend(["--cards", json.dumps(cards)])+ if brand:+ cmd.extend(["--brand", json.dumps(brand)])+ if font_family:+ cmd.extend(["--font-family", str(font_family)])if keep_caption_overlay:cmd.append("--keep-overlay")@@ -734,6 +768,11 @@ def generate_clip(allow_ass_fallback: bool = False,use_ass_captions: bool = False,keep_caption_overlay: bool = False,+ topic: Optional[dict] = None,+ progress: Optional[dict] = None,+ cards: Optional[list] = None,+ brand: Optional[dict] = None,+ font_family: Optional[str] = None,progress_callback: Optional[Callable[[int, str], None]] = None,) -> dict:"""@@ -1052,6 +1091,11 @@ def generate_clip(caption_font_scale=caption_font_scale,logo_position=logo_position,logo_scale=logo_scale,+ topic=topic,+ progress=progress,+ cards=cards,+ brand=brand,+ font_family=font_family,)if not remotion_ok and not allow_ass_fallback:@@ -1076,6 +1120,21 @@ def generate_clip(gradient_opacity = style_config.get("gradient_opacity", 0.6)use_logo = bool(logo_path)+ """+ The mark, the size Remotion would have drawn it.++ This path burns a logo with ffmpeg because Remotion could+ not run, and it used a flat 80px in a fixed corner: it read+ neither the frame height nor the template's logo scale. The+ same clip therefore carried a different mark depending on+ which renderer happened to succeed, and the studio preview+ only ever showed the Remotion one. These are Watermark.tsx's+ own numbers against the frame this render actually produced.+ """+ from services.video_processor import get_dimensions+ _, out_h = get_dimensions(cropped_path)+ logo_s = (out_h or REMOTION_REFERENCE_HEIGHT) / REMOTION_REFERENCE_HEIGHT+burn_captions(input_path=cropped_path,ass_path=ass_path,@@ -1083,9 +1142,9 @@ def generate_clip(gradient_overlay=use_gradient,gradient_opacity=gradient_opacity,logo_path=logo_path if use_logo else None,- logo_height=style_config.get("logo_height", 80),- logo_margin_x=style_config.get("logo_margin_x", 30),- logo_margin_y=style_config.get("logo_margin_y", 40),+ logo_height=round(REMOTION_LOGO_HEIGHT * logo_s * logo_scale),+ logo_margin_x=round(REMOTION_LOGO_EDGE * logo_s),+ logo_margin_y=round(REMOTION_LOGO_INSET * logo_s),logo_position=logo_position,)else:diff --git a/cli/VERSION b/cli/VERSIONindex 35b46ae..43c4dbe 100644--- a/cli/VERSION+++ b/cli/VERSION@@ -1 +1 @@-2.7.16+2.7.17diff --git a/package-lock.json b/package-lock.jsonindex d349279..d5f661a 100644--- a/package-lock.json+++ b/package-lock.json@@ -1,12 +1,12 @@{"name": "podcli",- "version": "2.7.16",+ "version": "2.7.17","lockfileVersion": 3,"requires": true,"packages": {"": {"name": "podcli",- "version": "2.7.16",+ "version": "2.7.17","license": "AGPL-3.0-only","dependencies": {"@fontsource/dm-sans": "^5.3.0",diff --git a/package.json b/package.jsonindex 744b649..ace545f 100644--- a/package.json+++ b/package.json@@ -1,6 +1,6 @@{"name": "podcli",- "version": "2.7.16",+ "version": "2.7.17","private": true,"description": "AI-powered podcast clip generator for TikTok/YouTube Shorts. Transcribe, find viral moments, export vertical clips with burned captions.","type": "module",diff --git a/remotion/render.mjs b/remotion/render.mjsindex a9ebe9a..17f3aa2 100644--- a/remotion/render.mjs+++ b/remotion/render.mjs@@ -12,6 +12,11 @@* --style branded \* --output /path/to/captioned.mp4 \* [--logo /path/to/logo.png] \+ * [--topic "Fitness"] [--topic-position top-left] \+ * [--progress] [--progress-color '#3B9CFF'] \+ * [--cards '[{"kind":"stat","start":2,"end":6,"value":"70%"}]'] \+ * [--brand '{"accent":"#4C9DF5","ink":"#FFFFFF","surface":"#0A0D14"}'] \+ * [--font-family "Inter"] \* [--fps 30]** node remotion/render.mjs --prebundle # Bundle only, no render@@ -25,7 +30,7 @@ import os from "os";import crypto from "crypto";-const BOOLEAN_FLAGS = new Set(["prebundle", "keep-overlay"]);+const BOOLEAN_FLAGS = new Set(["prebundle", "keep-overlay", "progress"]);function parseArgs() {const args = process.argv.slice(2);@@ -182,6 +187,38 @@ async function main() {}}+ /*+ * The parts a template switches on.+ *+ * Absent means null, and null draws nothing, so a caller that sends none of+ * these renders exactly what it rendered before they existed. A malformed+ * value is dropped with a line on stderr rather than failing the render: a+ * clip without its chip is worth more than no clip.+ */+ const json = (name) => {+ if (!opts[name]) return null;+ try {+ return JSON.parse(opts[name]);+ } catch {+ console.error(`Ignoring --${name}: not valid JSON`);+ return null;+ }+ };++ const cards = json("cards");+ const brand = json("brand");+ const topic = opts.topic+ ? {+ label: opts.topic,+ position: opts["topic-position"] || "top-left",+ ...(opts["topic-color"] ? { color: opts["topic-color"] } : {}),+ ...(opts["topic-background"] ? { background: opts["topic-background"] } : {}),+ }+ : null;+ const progress = opts.progress+ ? { ...(opts["progress-color"] ? { color: opts["progress-color"] } : {}) }+ : null;+const inputProps = {videoSrc,words,@@ -196,6 +233,11 @@ async function main() {captionFontScale: clampCaptionFontScale(opts["caption-font-scale"]),logoPosition: opts["logo-position"] || "top-left",logoScale: clampLogoScale(opts["logo-scale"]),+ topic,+ progress,+ cards: Array.isArray(cards) ? cards : null,+ brand: brand && typeof brand === "object" ? brand : null,+ fontFamily: opts["font-family"] || null,};console.log(diff --git a/remotion/src/CaptionedClip.tsx b/remotion/src/CaptionedClip.tsxindex 40f84d6..3e72717 100644--- a/remotion/src/CaptionedClip.tsx+++ b/remotion/src/CaptionedClip.tsx@@ -1,5 +1,5 @@import React from "react";-import { AbsoluteFill, useVideoConfig } from "remotion";+import { AbsoluteFill, useCurrentFrame, useVideoConfig } from "remotion";import { HormoziCaptions } from "./components/HormoziCaptions";import { KaraokeCaptions } from "./components/KaraokeCaptions";import { SubtleCaptions } from "./components/SubtleCaptions";@@ -7,6 +7,14 @@ import { BrandedCaptions } from "./components/BrandedCaptions";import { NameCard } from "./components/NameCard";import type { NameCardProps } from "./components/NameCard";import { Watermark } from "./components/Watermark";+import { TopicChip } from "./components/TopicChip";+import type { TopicChipProps } from "./components/TopicChip";+import { ProgressBar } from "./components/ProgressBar";+import type { ProgressBarProps } from "./components/ProgressBar";+import { Cards, CARD_CAPTION_MARGIN } from "./components/Cards";+import { cardAt } from "./cards";+import type { Card } from "./cards";+import type { Brand } from "./components/Cards";import { MOTION } from "./motion";import type { Motion } from "./motion";import type { Word, CaptionStyle, CaptionPosition, LogoPosition } from "./types";@@ -17,29 +25,78 @@ export interface CaptionedClipProps {style: CaptionStyle;logoSrc?: string;faceY?: number | null;- /** Who is speaking, shown for the first few seconds. */- nameCard?: NameCardProps | null;+ /** Face height as a fraction of the frame, when one was measured. */+ faceH?: number | null;captionPosition?: CaptionPosition;+ captionScale?: number;logoPosition?: LogoPosition;logoScale?: number;singleLine?: boolean;+ /** Who is speaking, shown for the first few seconds. */+ nameCard?: NameCardProps | null;+ /** A standing label saying what the clip is about. Null draws nothing. */+ topic?: TopicChipProps | null;+ /** How much of the clip is left. Null draws nothing. */+ progress?: ProgressBarProps | null;+ /** Cards that take the frame for a window each. Empty draws nothing. */+ cards?: Card[] | null;+ /** Where in the source file this composition begins, in frames. */+ startFrom?: number;+ /** The show's colours, for anything the renderer draws itself. */+ brand?: Brand | null;/** Per-part overrides; each part falls back to its style's own motion. */motion?: { captions?: Partial<Motion>; nameCard?: Partial<Motion> } | null;}export const CaptionedClip: React.FC<CaptionedClipProps> = ({+ videoSrc,words,style,logoSrc,faceY,- nameCard,+ faceH,captionPosition = "auto",+ captionScale: captionSize = 1,logoPosition = "top-left",logoScale = 1,singleLine = false,+ nameCard,+ topic,+ progress,+ cards,+ startFrom = 0,+ brand,motion,}) => {- const { height } = useVideoConfig();+ const { fps, height } = useVideoConfig();+ const frame = useCurrentFrame();++ /*+ * Captions drop toward the bottom edge while a card holds the frame.+ *+ * Their usual margin keeps them clear of a speaker's chin and hands. There+ * is no chin down there when a card is up, so the margin is only empty+ * surface, and it was the single biggest thing pushing the speaker's band+ * short enough to cut a face in half.+ */+ const cardUp = Boolean(cards?.length) && Boolean(cardAt(cards ?? [], frame / fps));+ const cardPlanned = Boolean(cards?.length);+ const captionShrink = cardUp ? 0.6 : cardPlanned ? 0.75 : 1;+ const baseCaptionStyle: CaptionStyle = cardUp+ ? { ...style, marginBottom: Math.min(style.marginBottom, CARD_CAPTION_MARGIN) }+ : style;+ // Keep a simple four-stop placement model. It is easier to reason about+ // than pixels, and stays proportional across every output shape.+ const placementMargin = captionPosition === "upper" ? 760+ : captionPosition === "center" ? 480+ : captionPosition === "lower" ? 220+ : baseCaptionStyle.marginBottom;+ const captionStyle: CaptionStyle = {+ ...baseCaptionStyle,+ marginBottom: placementMargin,+ fontSize: baseCaptionStyle.fontSize * captionSize * captionShrink,+ };+const captionMotion: Motion = {...(MOTION[style.name] ?? MOTION.subtle), ...(motion?.captions ?? {}),};@@ -53,14 +110,30 @@ export const CaptionedClip: React.FC<CaptionedClipProps> = ({return (<AbsoluteFill style={{ backgroundColor: "transparent" }}>+ {/* First, so everything below stays up while a card holds the frame. */}+ {cards && cards.length > 0 && (+ <Cards+ cards={cards}+ videoSrc={videoSrc}+ startFrom={startFrom}+ style={captionStyle}+ faceY={faceY}+ faceH={faceH}+ brand={brand}+ />+ )}<Watermark src={logoSrc} height={height} position={logoPosition} scale={logoScale} />{style.name === "branded" ? (- <BrandedCaptions words={words} style={style} logoSrc={logoSrc} faceY={faceY}- captionPosition={captionPosition} logoPosition={logoPosition} logoScale={logoScale} singleLine={singleLine} />+ <BrandedCaptions words={words} style={captionStyle} faceY={faceY}+ captionPosition={captionPosition} hasLogo={Boolean(logoSrc)}+ logoPosition={logoPosition} singleLine={singleLine} />) : (- <CaptionComponent words={words} style={style} singleLine={singleLine} motion={captionMotion} />+ <CaptionComponent words={words} style={captionStyle} motion={captionMotion}+ singleLine={singleLine} />)}{nameCard?.title && <NameCard {...nameCard} motion={cardMotion} />}+ {topic?.label && <TopicChip {...topic} />}+ {progress && <ProgressBar {...progress} />}</AbsoluteFill>);};diff --git a/remotion/src/Root.tsx b/remotion/src/Root.tsxindex 31a028e..12c442c 100644--- a/remotion/src/Root.tsx+++ b/remotion/src/Root.tsx@@ -3,9 +3,13 @@ import { Composition, continueRender, delayRender, getInputProps } from "remotioimport { CaptionedClip } from "./CaptionedClip";import { Bookend } from "./Bookend";import { Audiogram } from "./Audiogram";-import { STYLES } from "./types";+import { fontStack, STYLES } from "./types";import type { Word } from "./types";import type { CaptionPosition, LogoPosition } from "./types";+import type { TopicChipProps } from "./components/TopicChip";+import type { ProgressBarProps } from "./components/ProgressBar";+import type { Brand } from "./components/Cards";+import type { Card } from "./cards";import dmSans400 from "@fontsource/dm-sans/files/dm-sans-latin-400-normal.woff2";import dmSans700 from "@fontsource/dm-sans/files/dm-sans-latin-700-normal.woff2";@@ -65,6 +69,19 @@ const inputProps = getInputProps() as {captions?: Record<string, unknown>;nameCard?: Record<string, unknown>;} | null;+ /**+ * The parts a show's template switches on, and the colours it draws them in.+ *+ * Every one is optional and absent means "draw nothing", so a render that+ * sends none of them produces exactly the clip it produced before these+ * existed.+ */+ topic?: TopicChipProps | null;+ progress?: ProgressBarProps | null;+ cards?: Card[] | null;+ brand?: Brand | null;+ /** The show's own face, ahead of the stack that covers what it cannot draw. */+ fontFamily?: string | null;};export const RemotionRoot: React.FC = () => {@@ -82,6 +99,7 @@ export const RemotionRoot: React.FC = () => {...baseStyle,fontSize: baseStyle.fontSize * fontScale,marginBottom: positionMargins[captionPosition] ?? baseStyle.marginBottom,+ fontFamily: fontStack(inputProps.fontFamily),};return (@@ -105,6 +123,10 @@ export const RemotionRoot: React.FC = () => {logoScale: Math.max(0.5, Math.min(1.5, inputProps.logoScale || 1)),singleLine: inputProps.singleLine === true,motion: inputProps.motion ?? null,+ topic: inputProps.topic ?? null,+ progress: inputProps.progress ?? null,+ cards: inputProps.cards ?? null,+ brand: inputProps.brand ?? null,}}/><Compositiondiff --git a/remotion/src/cards.ts b/remotion/src/cards.tsnew file mode 100644index 0000000..c91c492--- /dev/null+++ b/remotion/src/cards.ts@@ -0,0 +1,118 @@+import type { Motion } from "./motion";++interface CardBase {+ /** Seconds on the clip's own clock. */+ start: number;+ end: number;+ accent?: string;+ background?: string;+ motion?: Motion;+ /**+ * Whether the speaker stays on screen above the card.+ *+ * Drawn unless a card says null. A podcast card that drops the person is the+ * news-explainer reflex this format is not: the moment is somebody saying+ * something, and cutting their face out mid-sentence throws away the only+ * thing the clip has.+ *+ * Null is for the card that earns the whole frame, like a screenshot nobody+ * could read at half the height.+ */+ speaker?: null;+}++/**+ * What a card can say.+ *+ * Four kinds rather than a free-form layout, because a card is only worth+ * taking the frame for when it says something the talking head cannot: a+ * number, a claim, a list, or the sentence itself set large enough to read+ * at arm's length. Anything past that is a slide deck.+ */+export type Card =+ | (CardBase & {+ kind: "stat";+ /** The number, and only the number. */+ value: string;+ eyebrow?: string;+ caption?: string;+ })+ | (CardBase & {+ kind: "headline";+ eyebrow?: string;+ lead: string;+ /** The payload word, set apart from the lead. */+ emphasis?: string;+ sub?: string;+ })+ | (CardBase & { kind: "bullets"; eyebrow?: string; items: string[] })+ /**+ * Two or three things measured against each other.+ *+ * Emphasis rather than a colour per bar: the story is that one of them is+ * the point, and giving every bar its own hue buries the one that is. The+ * subject wears the accent, the rest are context.+ */+ | (CardBase & {+ kind: "compare";+ eyebrow?: string;+ rows: { label: string; value: number; display?: string; subject?: boolean }[];+ })+ /** One thing, before and after. A dumbbell, in two steps of one hue. */+ | (CardBase & {+ kind: "change";+ eyebrow?: string;+ label?: string;+ from: { value: string; note?: string };+ to: { value: string; note?: string };+ })+ /** One ratio against its whole. A meter, never a two-slice pie. */+ | (CardBase & {+ kind: "share";+ eyebrow?: string;+ /** 0 to 1. */+ value: number;+ display?: string;+ caption?: string;+ })+ /** A name nobody in the audience necessarily knows, with its mark. */+ | (CardBase & {+ kind: "entity";+ eyebrow?: string;+ name: string;+ note?: string;+ /** Logo or photograph. A URL the renderer can reach, or a bundle path. */+ src?: string;+ })+ | (CardBase & { kind: "quote"; text: string; attribution?: string })+ | (CardBase & {+ kind: "image";+ /** A URL the renderer can reach, or a path inside the bundle. */+ src: string;+ /**+ * "fit" shows the whole picture, which is what a screenshot needs: a+ * cropped headline is not evidence of anything. "fill" covers the frame+ * and is for a photograph, where the edges carry nothing.+ */+ fit?: "fit" | "fill";+ caption?: string;+ });++/**+ * The card in front at this moment, or none.+ *+ * A later card wins an overlap rather than the array order deciding it,+ * because the planner writes cards in the order they were found and two+ * windows that touch would otherwise flicker between them frame by frame.+ * The window is half-open so a card ending where the next begins is a cut,+ * not a frame showing both.+ */+export const cardAt = (cards: Card[], seconds: number): Card | null => {+ let found: Card | null = null;+ for (const card of cards) {+ if (seconds >= card.start && seconds < card.end) {+ if (!found || card.start >= found.start) found = card;+ }+ }+ return found;+};diff --git a/remotion/src/components/BrandedCaptions.tsx b/remotion/src/components/BrandedCaptions.tsxindex 8cd70c1..966d61b 100644--- a/remotion/src/components/BrandedCaptions.tsx+++ b/remotion/src/components/BrandedCaptions.tsx@@ -5,20 +5,24 @@ import {spring,} from "remotion";import type { Word, CaptionStyle, CaptionPosition, LogoPosition } from "../types";-import { captionScale } from "../types";+import {+ captionScale, LOGO_CAPTION_GAP, LOGO_HEIGHT, LOGO_INSET,+} from "../types";import { buildChunks, activeChunkAt, splitCaptionLines } from "../chunks";-import { LOGO_CAPTION_GAP, LOGO_HEIGHT, LOGO_INSET } from "./Watermark";interface Props {words: Word[];style: CaptionStyle;- // The logo is drawn by Watermark now, but the caption margin still has to- // know it is there.- logoSrc?: string;faceY?: number | null; // normalized 0-1 (0=top, 1=bottom)captionPosition?: CaptionPosition;+ /**+ * Whether a logo is on the frame, and where. The mark itself is drawn by+ * Watermark one level up so all four styles carry it; this component still+ * has to know, because a bottom-anchored logo and a low caption want the+ * same band of pixels.+ */+ hasLogo?: boolean;logoPosition?: LogoPosition;- logoScale?: number;singleLine?: boolean;}@@ -33,7 +37,8 @@ const WordWithPill: React.FC<{isActive: boolean;frame: number;fps: number;-}> = ({ word, isActive, frame, fps }) => {+ emphasisColor?: string;+}> = ({ word, isActive, frame, fps, emphasisColor }) => {const { height } = useVideoConfig();const s = captionScale(height);const wordEntryFrame = Math.round(word.start * fps);@@ -64,7 +69,15 @@ const WordWithPill: React.FC<{}}/>{/* Word text — always inline, never shifts */}- <span style={{ position: "relative", zIndex: 1 }}>{word.word}</span>+ <span+ style={{+ position: "relative",+ zIndex: 1,+ ...(word.emphasis && emphasisColor ? { color: emphasisColor } : {}),+ }}+ >+ {word.word}+ </span></span>);};@@ -106,6 +119,7 @@ const CaptionLine: React.FC<{isActive={isActive}frame={frame}fps={fps}+ emphasisColor={style.emphasisColor}/></React.Fragment>);@@ -117,11 +131,10 @@ const CaptionLine: React.FC<{export const BrandedCaptions: React.FC<Props> = ({words,style,- logoSrc,faceY,captionPosition = "auto",+ hasLogo = false,logoPosition = "top-left",- logoScale = 1,singleLine = false,}) => {const frame = useCurrentFrame();@@ -152,8 +165,8 @@ export const BrandedCaptions: React.FC<Props> = ({}// A bottom-anchored logo spans 180-306 scaled units. Captions sitting inside// that band (captionPosition "lower" starts at 220) would render over it.- if (logoSrc && logoPosition.startsWith("bottom-")) {- dynamicMargin = Math.max(dynamicMargin, (LOGO_INSET + LOGO_HEIGHT * logoScale + LOGO_CAPTION_GAP) * s);+ if (hasLogo && logoPosition.startsWith("bottom-")) {+ dynamicMargin = Math.max(dynamicMargin, (LOGO_INSET + LOGO_HEIGHT + LOGO_CAPTION_GAP) * s);}return (diff --git a/remotion/src/components/Cards.tsx b/remotion/src/components/Cards.tsxnew file mode 100644index 0000000..ca94716--- /dev/null+++ b/remotion/src/components/Cards.tsx@@ -0,0 +1,642 @@+import React from "react";+import {+ getRemotionEnvironment, Img, OffthreadVideo, Video, staticFile,+ useCurrentFrame, useVideoConfig,+} from "remotion";+import { captionScale, FONT } from "../types";+import type { CaptionStyle } from "../types";+import { MOTION, motionAt } from "../motion";+import { cardAt } from "../cards";+import type { Card } from "../cards";++/**+ * A card keeps the person on screen.+ *+ * The formats this borrows from are news explainers: they cut away from the+ * presenter constantly, because what they are narrating is an announcement and+ * the presenter is only reading it out. A podcast is the other way round. The+ * moment is somebody saying something, and cutting their face out mid-sentence+ * throws away the only thing the clip has.+ *+ * So a card takes the band directly above the captions and the speaker holds+ * everything above that. The eye travels from face to point to words without+ * ever leaving the frame, and a card that arrives is an addition rather than a+ * cutaway to a different video.+ */++/**+ * The show's colours, or ours when it has not set any.+ *+ * Three values rather than a palette: the one colour that means "this is the+ * thing", what text is set in, and what it is set on. Everything else on a+ * card is one of those three at a lower opacity, which is what keeps a card+ * looking like the show rather than like a theme.+ */+export type Brand = { accent: string; ink: string; surface: string };++export const DEFAULT_BRAND: Brand = {+ accent: "#4C9DF5",+ ink: "#FFFFFF",+ surface: "#0A0D14",+};++/** Text that is not the point, and marks that are not the subject. */+const muted = (ink: string) => `color-mix(in oklab, ${ink} 62%, transparent)`;+const context = (ink: string) => `color-mix(in oklab, ${ink} 22%, transparent)`;++/**+ * A three-step rhythm, so grouping reads from the gaps rather than from rules.+ * The first pass used 22, 24, 26, 28 and 32, which is one gap five times.+ */+const GAP = { tight: 14, group: 36, section: 76 };++/**+ * Marks, thin and unfussy.+ *+ * `cap` rounds the data end of a bar only. `gap` is the surface showing between+ * adjacent marks, which is what keeps two bars from reading as one shape.+ */+const MARK = { bar: 26, cap: 8, gap: 30, rule: 5, dot: 26 };++/** A meter's unfilled remainder: the fill's own hue, several steps lighter. */+const track = (fill: string) =>+ `color-mix(in oklab, ${fill} 22%, transparent)`;++/** Authored for the 1920-tall canvas, like every other measurement here. */+const TYPE = { label: 34, body: 46, item: 58, quote: 68, lead: 84, figure: 168 };++/**+ * The least of the frame the speaker keeps.+ *+ * The band is whatever the card does not need rather than a fixed height: a+ * two-line quote and a bare figure want different amounts, and pinning both+ * ends meant a tall card overflowed upward across the speaker's chin. This+ * floor stops a very long card from squeezing the face out altogether.+ */+const SPEAKER_MIN = 820;++/**+ * How much room a head needs, given how big the face is.+ *+ * A detected face box is the face, not the head: it stops at the hairline and+ * at the chin, and a band cut to it lands exactly on both. Two and a half+ * times leaves the skull above and the jaw and some shoulder below, which is+ * the framing anyone would choose by hand.+ *+ * Without a measurement this falls back to the flat floor above, which is a+ * guess that happens to suit a mid-shot and is wrong on a close-up. That is+ * the honest state of it: the detector has the number, an older engine does+ * not send it, and a clip rendered by one should not look broken.+ */+const HEAD_TO_FACE = 2.5;++/**+ * Where captions sit while a card is up.+ *+ * Their usual margin exists to clear a speaker's chin and hands, and there is+ * no chin down there behind a card. Dropping them toward the bottom edge is+ * what buys the speaker a band tall enough to hold a whole head.+ */+export const CARD_CAPTION_MARGIN = 180;++/**+ * The room the captions need, measured rather than assumed.+ *+ * A fixed number worked until a two-line Georgian caption grew taller than it+ * and the card's last line rendered behind the pill. The caption's own style+ * knows where it sits and how big it is, so the reserved band is read off that:+ * its margin, two lines of it, and a gap.+ */+const captionZone = (style: CaptionStyle) =>+ style.marginBottom + style.fontSize * CAPTION_LINES * 1.2 + GAP.group;++/**+ * How many lines of caption a card gets out of the way of.+ *+ * Two was the common case and wrong at the worst one: a chunk that wrapped to+ * three lines covered the card's last row, which is the row a comparison keeps+ * its second bar in. Reserving for the tallest caption costs a card some+ * height on every clip and costs it nothing on the clip where it matters.+ */+const CAPTION_LINES = 3;++/*+ * The panel runs to the bottom edge, and that is not negotiable.+ *+ * It was ended under the caption for a while, to close the empty strip a+ * caption's own margin leaves below it. What that actually did was put a band+ * of the speaker back under the card: head above, panel across, chin below,+ * so one person appeared twice in a frame with a slab through the middle of+ * their face. Empty surface under a caption reads as a floor. A face cut in+ * half reads as a mistake, and it is the more expensive of the two by far.+ */++/**+ * The line above the payload, when there is something worth saying there.+ *+ * Set in the same white as everything else rather than the accent. An accent+ * on the label and on the figure is the same swatch twice on one card, which+ * is how a colour stops meaning anything.+ */+const Label: React.FC<{ text: string; scale: number; tone: string }> = ({+ text, scale: s, tone,+}) => (+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.label * s,+ fontWeight: 600,+ letterSpacing: 1.5 * s,+ color: tone,+ marginBottom: GAP.tight * s,+ }}+ >+ {text}+ </div>+);++/** One end of a change, as a value with its note under it. */+const Endpoint: React.FC<{+ value: string; note?: string; tone: string; scale: number; brand: Brand;+}> = ({ value, note, tone, scale: s, brand }) => (+ <div style={{ textAlign: "center" }}>+ <div+ style={{+ width: MARK.dot * s, height: MARK.dot * s, borderRadius: "50%",+ backgroundColor: tone, margin: "0 auto",+ }}+ />+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.item * s, fontWeight: 700,+ color: brand.ink, marginTop: 12 * s, lineHeight: 1,+ }}+ >+ {value}+ </div>+ {note && (+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.label * s, color: muted(brand.ink),+ marginTop: 6 * s,+ }}+ >+ {note}+ </div>+ )}+ </div>+);++const CardBody: React.FC<{ card: Card; scale: number; brand: Brand }> = ({+ card, scale: s, brand,+}) => {+ const accent = card.accent ?? brand.accent;+ const INK = brand.ink;+ const MUTED = muted(brand.ink);+ const CONTEXT = context(brand.ink);++ if (card.kind === "stat") {+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.figure * s,+ fontWeight: 700,+ lineHeight: 0.95,+ letterSpacing: -4 * s,+ color: accent,+ }}+ >+ {card.value}+ </div>+ {card.caption && (+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.body * s,+ fontWeight: 400,+ lineHeight: 1.3,+ color: INK,+ marginTop: GAP.tight * s,+ maxWidth: "80%",+ }}+ >+ {card.caption}+ </div>+ )}+ </>+ );+ }++ if (card.kind === "headline") {+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.lead * s,+ fontWeight: 600,+ lineHeight: 1.12,+ letterSpacing: -1 * s,+ color: INK,+ }}+ >+ {card.lead}+ {card.emphasis && (+ // The payload word carries the card's one accent. Italic as well+ // would be two ways of saying the same thing.+ <span style={{ color: accent }}> {card.emphasis}</span>+ )}+ </div>+ {card.sub && (+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.body * s,+ fontWeight: 400,+ lineHeight: 1.3,+ color: MUTED,+ marginTop: GAP.tight * s,+ maxWidth: "80%",+ }}+ >+ {card.sub}+ </div>+ )}+ </>+ );+ }++ if (card.kind === "bullets") {+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ {card.items.map((item, i) => (+ <div+ key={i}+ style={{+ display: "flex",+ alignItems: "center",+ gap: 20 * s,+ marginTop: i === 0 ? 0 : GAP.tight * s,+ }}+ >+ {/* A rule rather than a bullet glyph or an icon: it is a list+ marker, and a dot in a coloured circle is chrome. */}+ <div+ style={{+ width: 26 * s,+ height: 3 * s,+ backgroundColor: i === 0 ? accent : MUTED,+ flexShrink: 0,+ }}+ />+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.item * s,+ fontWeight: 500,+ lineHeight: 1.25,+ color: INK,+ }}+ >+ {item}+ </div>+ </div>+ ))}+ </>+ );+ }++ if (card.kind === "compare") {+ const top = Math.max(...card.rows.map((r) => r.value)) || 1;+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ {card.rows.slice(0, 3).map((row, i) => (+ <div key={i} style={{ marginTop: i === 0 ? 0 : MARK.gap * s }}>+ <div+ style={{+ display: "flex",+ justifyContent: "space-between",+ alignItems: "baseline",+ marginBottom: 8 * s,+ }}+ >+ {/* Ink, not the bar's colour: the mark beside it carries identity. */}+ <span style={{ fontFamily: FONT, fontSize: TYPE.label * s, color: MUTED }}>+ {row.label}+ </span>+ <span+ style={{+ fontFamily: FONT, fontSize: TYPE.body * s, fontWeight: 700,+ color: row.subject ? INK : MUTED,+ }}+ >+ {row.display ?? row.value}+ </span>+ </div>+ <div+ style={{+ width: `${Math.max(4, (row.value / top) * 100)}%`,+ height: MARK.bar * s,+ backgroundColor: row.subject ? accent : CONTEXT,+ // Rounded at the data end only. A bar rounded at the baseline+ // reads as floating away from the axis it is measured from.+ borderRadius: `0 ${MARK.cap * s}px ${MARK.cap * s}px 0`,+ }}+ />+ </div>+ ))}+ </>+ );+ }++ if (card.kind === "change") {+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ <div style={{ display: "flex", alignItems: "center", gap: 0 }}>+ <Endpoint value={card.from.value} note={card.from.note} tone={CONTEXT} scale={s} brand={brand} />+ <div+ style={{+ flex: 1,+ height: MARK.rule * s,+ // One hue in two steps: the same measure, later.+ background: `linear-gradient(90deg, ${CONTEXT}, ${accent})`,+ margin: `0 ${18 * s}px`,+ transform: `translateY(${-14 * s}px)`,+ }}+ />+ <Endpoint value={card.to.value} note={card.to.note} tone={accent} scale={s} brand={brand} />+ </div>+ </>+ );+ }++ if (card.kind === "share") {+ const pct = Math.max(0, Math.min(1, card.value));+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.figure * 0.72 * s, fontWeight: 700,+ lineHeight: 1, letterSpacing: -3 * s, color: accent,+ marginBottom: GAP.tight * s,+ }}+ >+ {card.display ?? `${Math.round(pct * 100)}%`}+ </div>+ {/* The unfilled track is a lighter step of the fill's own ramp, so the+ whole bar reads as one measure rather than as bar-on-background. */}+ <div+ style={{+ width: "100%", height: MARK.bar * s,+ backgroundColor: track(accent), borderRadius: MARK.cap * s,+ overflow: "hidden",+ }}+ >+ <div style={{ width: `${pct * 100}%`, height: "100%", backgroundColor: accent }} />+ </div>+ {card.caption && (+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.body * s, color: INK,+ lineHeight: 1.3, marginTop: GAP.tight * s, maxWidth: "85%",+ }}+ >+ {card.caption}+ </div>+ )}+ </>+ );+ }++ if (card.kind === "entity") {+ return (+ <>+ {card.eyebrow && <Label text={card.eyebrow} scale={s} tone={MUTED} />}+ <div style={{ display: "flex", alignItems: "center", gap: 28 * s }}>+ {card.src && (+ <Img+ src={card.src.startsWith("http") ? card.src : staticFile(card.src)}+ style={{+ width: 132 * s, height: 132 * s, objectFit: "cover",+ borderRadius: 20 * s, flexShrink: 0,+ outline: `${1 * s}px solid rgba(255,255,255,0.1)`,+ outlineOffset: `${-1 * s}px`,+ }}+ />+ )}+ <div>+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.item * s, fontWeight: 700,+ lineHeight: 1.15, color: INK,+ }}+ >+ {card.name}+ </div>+ {card.note && (+ <div+ style={{+ fontFamily: FONT, fontSize: TYPE.body * s, color: MUTED,+ lineHeight: 1.3, marginTop: 6 * s,+ }}+ >+ {card.note}+ </div>+ )}+ </div>+ </div>+ </>+ );+ }++ if (card.kind === "image") {+ return (+ <>+ <Img+ src={card.src.startsWith("http") ? card.src : staticFile(card.src)}+ style={{+ width: "100%",+ maxHeight: 520 * s,+ objectFit: card.fit === "fill" ? "cover" : "contain",+ borderRadius: 16 * s,+ // A screenshot's own edge is often near the surface tone behind it.+ // Pure white at a tenth, never a tone borrowed from the palette.+ outline: `${1 * s}px solid rgba(255,255,255,0.1)`,+ outlineOffset: `${-1 * s}px`,+ }}+ />+ {card.caption && (+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.label * s,+ fontWeight: 400,+ lineHeight: 1.3,+ color: MUTED,+ marginTop: GAP.tight * s,+ }}+ >+ {card.caption}+ </div>+ )}+ </>+ );+ }++ return (+ <>+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.quote * s,+ fontWeight: 500,+ lineHeight: 1.25,+ letterSpacing: -0.5 * s,+ color: INK,+ // Hung, so the first letter of the line sits on the same edge as+ // everything below it rather than a quote mark's worth to the right.+ textIndent: `${-18 * s}px`,+ }}+ >+ “{card.text}”+ </div>+ {card.attribution && (+ <div+ style={{+ fontFamily: FONT,+ fontSize: TYPE.label * s,+ fontWeight: 600,+ letterSpacing: 1.5 * s,+ color: MUTED,+ marginTop: GAP.group * s,+ }}+ >+ {card.attribution}+ </div>+ )}+ </>+ );+};++/**+ * The speaker, still on screen, above the card.+ *+ * Each environment gets the video component it can actually play. The+ * in-browser player seeks a <video> element per frame, and OffthreadVideo+ * hangs there when the file is served without range requests, which left every+ * preview stuck on the first card. A render is the reverse: it has no seekable+ * player, and a <video> element it has to drive frame by frame stalls until+ * the render times out with nothing drawn.+ */+const SpeakerBand: React.FC<{ src: string; startFrom: number; faceY?: number | null }> = ({+ src, startFrom, faceY,+}) => {+ const Frame = getRemotionEnvironment().isRendering ? OffthreadVideo : Video;+ return (+ <Frame+ src={src}+ startFrom={startFrom}+ muted+ style={{+ width: "100%",+ height: "100%",+ objectFit: "cover",+ objectPosition: `50% ${(faceY ?? 0.4) * 100}%`,+ }}+ />+ );+};++export const Cards: React.FC<{+ cards: Card[];+ videoSrc?: string;+ startFrom?: number;+ style: CaptionStyle;+ faceY?: number | null;+ faceH?: number | null;+ brand?: Brand | null;+}> = ({ cards, videoSrc, startFrom = 0, style, faceY, faceH, brand }) => {+ const colours = { ...DEFAULT_BRAND, ...(brand ?? {}) };+ const frame = useCurrentFrame();+ const { fps, height } = useVideoConfig();+ const s = captionScale(height);++ // A picture whose file went missing is dropped here rather than drawn: the+ // card is opaque, so rendering it anyway would cover the clip with nothing.+ const usable = cards.filter((c) => c.kind !== "image" || Boolean(c.src));+ const card = cardAt(usable, frame / fps);+ if (!card) return null;++ const { opacity } = motionAt({+ frame, fps,+ start: card.start,+ end: card.end,+ motion: card.motion ?? MOTION.card,+ });+ if (opacity <= 0) return null;++ /*+ * The speaker stays, unless what is left of the frame would cut them in half.+ *+ * A band shorter than this shows a forehead and a mouth, which reads worse+ * than not showing the person at all: a viewer reads a slice of a face as a+ * mistake and a full-frame card as a decision. So it is a whole head or+ * none, never part of one.+ */+ const room = height - (captionZone(style) + GAP.section) * s;+ const headNeeds = faceH ? faceH * HEAD_TO_FACE * height : SPEAKER_MIN * s;+ const withSpeaker = card.speaker !== null+ && Boolean(videoSrc)+ && room >= headNeeds;++ return (+ <div+ style={{+ position: "absolute",+ inset: 0,+ backgroundColor: card.background ?? colours.surface,+ opacity,+ display: "flex",+ flexDirection: "column",+ // The captions are not part of this stack; they are drawn over it.+ // Reserving their band as padding is what keeps a card's last line+ // from ending up behind the pill.+ paddingBottom: captionZone(style) * s,+ }}+ >+ {withSpeaker && videoSrc && (+ <div style={{ flex: 1, minHeight: SPEAKER_MIN * s, overflow: "hidden" }}>+ <SpeakerBand src={videoSrc} startFrom={startFrom} faceY={faceY} />+ </div>+ )}+ <div+ style={{+ padding: `${GAP.section * s}px ${96 * s}px 0`,+ display: "flex",+ flexDirection: "column",+ /*+ * Under the speaker when there is one, centred when there is not.+ *+ * A card that has the whole frame and puts its content against the+ * captions leaves two thirds of the frame empty above it, which+ * reads as a card that failed to load. With a speaker the content+ * belongs low, because the thing above it is a person and the gap+ * between them is what makes the pair read as one shot.+ */+ marginTop: "auto",+ marginBottom: withSpeaker ? undefined : "auto",+ }}+ >+ <CardBody card={card} scale={s} brand={colours} />+ </div>+ </div>+ );+};diff --git a/remotion/src/components/HormoziCaptions.tsx b/remotion/src/components/HormoziCaptions.tsxindex cb38062..89214ab 100644--- a/remotion/src/components/HormoziCaptions.tsx+++ b/remotion/src/components/HormoziCaptions.tsx@@ -16,7 +16,9 @@ interface Props {motion?: Motion;}-export const HormoziCaptions: React.FC<Props> = ({ words, style, singleLine = false, motion }) => {+export const HormoziCaptions: React.FC<Props> = ({+ words, style, singleLine = false, motion,+}) => {const frame = useCurrentFrame();const { fps, height, durationInFrames } = useVideoConfig();const s = captionScale(height);@@ -73,13 +75,19 @@ export const HormoziCaptions: React.FC<Props> = ({ words, style, singleLine = faconst text = style.uppercase? word.word.toUpperCase(): word.word;+ // The word being spoken still wins: the sweep is what this style is.+ // Emphasis colours it for the rest of the chunk, which is the part+ // somebody reads with the sound off.+ const emphasis = word.emphasis+ ? style.emphasisColor ?? style.activeColor+ : null;return (<React.Fragment key={i}>{i > 0 ? " " : ""}<spanstyle={{- color: isActive ? style.activeColor : style.color,+ color: isActive ? style.activeColor : emphasis ?? style.color,textShadow: isActive? `0 0 30px ${style.activeColor}60, 0 0 60px ${style.activeColor}30`: "none",diff --git a/remotion/src/components/KaraokeCaptions.tsx b/remotion/src/components/KaraokeCaptions.tsxindex c45f274..1f76e9c 100644--- a/remotion/src/components/KaraokeCaptions.tsx+++ b/remotion/src/components/KaraokeCaptions.tsx@@ -40,8 +40,17 @@ const KaraokeLine: React.FC<{<React.Fragment key={i}>{i > 0 ? " " : ""}<span style={{ position: "relative", display: "inline" }}>- {/* Base text (dim) */}- <span style={{ color: style.color }}>{word.word}</span>+ {/* Base text (dim). Emphasis lifts it out of the dim state so a+ key word is legible before the sweep reaches it. */}+ <span+ style={{+ color: word.emphasis+ ? style.emphasisColor ?? style.activeColor+ : style.color,+ }}+ >+ {word.word}+ </span>{/* Active overlay (clips left to right) */}<spanstyle={{diff --git a/remotion/src/components/NameCard.tsx b/remotion/src/components/NameCard.tsxindex bee1a10..9dcdfa0 100644--- a/remotion/src/components/NameCard.tsx+++ b/remotion/src/components/NameCard.tsx@@ -1,6 +1,6 @@import React from "react";import { useCurrentFrame, useVideoConfig } from "remotion";-import { captionScale } from "../types";+import { captionScale, FONT } from "../types";import { MOTION, motionAt } from "../motion";import type { Motion } from "../motion";@@ -40,7 +40,6 @@ export const NameCard: React.FC<NameCardProps> = ({const s = captionScale(height);if (!title) return null;- if (frame >= seconds * fps) return null;const { opacity, shift } = motionAt({frame, fps, start: 0, end: seconds,@@ -67,7 +66,7 @@ export const NameCard: React.FC<NameCardProps> = ({><divstyle={{- fontFamily: "'DM Sans', sans-serif",+ fontFamily: FONT,fontSize: 44 * s,fontWeight: 700,lineHeight: 1.2,@@ -79,7 +78,7 @@ export const NameCard: React.FC<NameCardProps> = ({{subtitle && (<divstyle={{- fontFamily: "'DM Sans', sans-serif",+ fontFamily: FONT,fontSize: 38 * s,fontWeight: 400,lineHeight: 1.25,diff --git a/remotion/src/components/ProgressBar.tsx b/remotion/src/components/ProgressBar.tsxnew file mode 100644index 0000000..dcd80d7--- /dev/null+++ b/remotion/src/components/ProgressBar.tsx@@ -0,0 +1,58 @@+import React from "react";+import { useCurrentFrame, useVideoConfig } from "remotion";+import { captionScale } from "../types";++export interface ProgressBarProps {+ color?: string;+ /** The unfilled remainder. Transparent leaves the video showing through. */+ track?: string;+ /** Bar thickness in unscaled units. */+ thickness?: number;+}++/**+ * How much of the clip is left, as a line across the bottom.+ *+ * A viewer deciding whether to keep watching is asking how long this goes on+ * for, and a clip that answers keeps them past the first seconds. The player's+ * own scrubber does not: it belongs to the feed, it is not always drawn, and+ * on a reposted file it is gone.+ *+ * Anchored to the very bottom edge rather than inset, because a progress bar+ * that floats reads as a design element and one on the edge reads as a fact.+ */+export const ProgressBar: React.FC<ProgressBarProps> = ({+ color = "#3B9CFF",+ track = "rgba(255,255,255,0.15)",+ thickness = 8,+}) => {+ const frame = useCurrentFrame();+ const { durationInFrames, height } = useVideoConfig();+ const s = captionScale(height);++ // The last frame should read as full rather than as one frame short.+ const progress = durationInFrames > 1+ ? Math.min(1, frame / (durationInFrames - 1))+ : 1;++ return (+ <div+ style={{+ position: "absolute",+ left: 0,+ right: 0,+ bottom: 0,+ height: thickness * s,+ backgroundColor: track,+ }}+ >+ <div+ style={{+ width: `${progress * 100}%`,+ height: "100%",+ backgroundColor: color,+ }}+ />+ </div>+ );+};diff --git a/remotion/src/components/SubtleCaptions.tsx b/remotion/src/components/SubtleCaptions.tsxindex 9c4b7a4..3997ea9 100644--- a/remotion/src/components/SubtleCaptions.tsx+++ b/remotion/src/components/SubtleCaptions.tsx@@ -13,7 +13,51 @@ interface Props {motion?: Motion;}-export const SubtleCaptions: React.FC<Props> = ({ words, style, singleLine = false, motion }) => {+/**+ * One line of the caption, word by word.+ *+ * Drawn as spans rather than as a joined string so an emphasised word can be+ * coloured; the line reads identically when nothing is emphasised.+ */+const Line: React.FC<{+ words: Word[];+ style: CaptionStyle;+ scale: number;+ nowrap?: boolean;+}> = ({ words, style, scale, nowrap = false }) => (+ <span+ style={{+ fontFamily: style.fontFamily,+ fontSize: style.fontSize * scale,+ fontWeight: 400,+ color: style.color,+ textShadow:+ "0 1px 3px rgba(0,0,0,0.95), 0 0 20px rgba(0,0,0,0.6), 0 0 50px rgba(0,0,0,0.3)",+ textAlign: "center",+ lineHeight: 1.35,+ whiteSpace: nowrap ? "nowrap" : undefined,+ }}+ >+ {words.map((word, i) => (+ <React.Fragment key={i}>+ {i > 0 ? " " : ""}+ <span+ style={+ word.emphasis+ ? { color: style.emphasisColor ?? style.activeColor }+ : undefined+ }+ >+ {word.word}+ </span>+ </React.Fragment>+ ))}+ </span>+);++export const SubtleCaptions: React.FC<Props> = ({+ words, style, singleLine = false, motion,+}) => {const frame = useCurrentFrame();const { fps, height, durationInFrames } = useVideoConfig();const s = captionScale(height);@@ -41,9 +85,6 @@ export const SubtleCaptions: React.FC<Props> = ({ words, style, singleLine = falMath.ceil(activeChunk.words.length / 2),singleLine,);- const text1 = line1.map((w) => w.word).join(" ");- const text2 = line2.map((w) => w.word).join(" ");-return (<divstyle={{@@ -59,37 +100,8 @@ export const SubtleCaptions: React.FC<Props> = ({ words, style, singleLine = faltransform: `translateY(${translateY}px)`,}}>- <span- style={{- fontFamily: style.fontFamily,- fontSize: style.fontSize * s,- fontWeight: 400,- color: style.color,- textShadow:- "0 1px 3px rgba(0,0,0,0.95), 0 0 20px rgba(0,0,0,0.6), 0 0 50px rgba(0,0,0,0.3)",- textAlign: "center",- lineHeight: 1.35,- whiteSpace: singleLine ? "nowrap" : undefined,- }}- >- {text1}- </span>- {text2 && (- <span- style={{- fontFamily: style.fontFamily,- fontSize: style.fontSize * s,- fontWeight: 400,- color: style.color,- textShadow:- "0 1px 3px rgba(0,0,0,0.95), 0 0 20px rgba(0,0,0,0.6), 0 0 50px rgba(0,0,0,0.3)",- textAlign: "center",- lineHeight: 1.35,- }}- >- {text2}- </span>- )}+ <Line words={line1} style={style} scale={s} nowrap={singleLine} />+ {line2.length > 0 && <Line words={line2} style={style} scale={s} />}</div>);};diff --git a/remotion/src/components/TopicChip.tsx b/remotion/src/components/TopicChip.tsxnew file mode 100644index 0000000..4446005--- /dev/null+++ b/remotion/src/components/TopicChip.tsx@@ -0,0 +1,67 @@+import React from "react";+import { useVideoConfig } from "remotion";+import { captionScale, FONT, LOGO_EDGE, LOGO_INSET } from "../types";+import type { LogoPosition } from "../types";++export interface TopicChipProps {+ /** What the clip is about, in two or three words. */+ label: string;+ /** Where it sits. Shares LogoPosition so a template names corners once. */+ position?: LogoPosition;+ color?: string;+ /** Drawn behind the label. Transparent by default, which is the reference look. */+ background?: string;+}++/**+ * A standing label saying what this clip is about.+ *+ * Every produced short carries one and no cut of a podcast does, which is most+ * of why a raw clip reads as a raw clip. It holds still for the whole clip on+ * purpose: it is furniture, not a caption, and anything that moves up there+ * competes with the words.+ *+ * Small, letterspaced and upper case, so it reads as a label rather than as a+ * sentence somebody forgot to finish.+ */+export const TopicChip: React.FC<TopicChipProps> = ({+ label,+ position = "top-left",+ color = "#FFFFFF",+ background = "transparent",+}) => {+ const { height } = useVideoConfig();+ const s = captionScale(height);++ if (!label) return null;++ const padded = background !== "transparent";++ return (+ <div+ style={{+ position: "absolute",+ ...(position.startsWith("top-")+ ? { top: LOGO_INSET * s }+ : { bottom: LOGO_INSET * s }),+ ...(position.endsWith("-left")+ ? { left: LOGO_EDGE * s }+ : position.endsWith("-right")+ ? { right: LOGO_EDGE * s }+ : { left: "50%", transform: "translateX(-50%)" }),+ fontFamily: FONT,+ fontSize: 30 * s,+ fontWeight: 700,+ letterSpacing: 3 * s,+ textTransform: "uppercase",+ color,+ background,+ ...(padded+ ? { padding: `${10 * s}px ${20 * s}px`, borderRadius: 10 * s }+ : { textShadow: "0 2px 12px rgba(0,0,0,0.8)" }),+ }}+ >+ {label}+ </div>+ );+};diff --git a/remotion/src/components/Watermark.tsx b/remotion/src/components/Watermark.tsxindex c4c9ba4..7ea5caa 100644--- a/remotion/src/components/Watermark.tsx+++ b/remotion/src/components/Watermark.tsx@@ -1,13 +1,7 @@import React from "react";import { Img, staticFile } from "remotion";+import { captionScale, LOGO_EDGE, LOGO_HEIGHT, LOGO_INSET, LOGO_WIDTH } from "../types";import type { LogoPosition } from "../types";-import { captionScale } from "../types";--// Unscaled logo box, shared with BrandedCaptions' caption-margin guard so the-// mark and the gap left for it cannot drift apart.-export const LOGO_INSET = 180;-export const LOGO_HEIGHT = 126;-export const LOGO_CAPTION_GAP = 24;/*** The show's logo, on every clip.@@ -22,7 +16,6 @@ export const Watermark: React.FC<{src?: string;height: number;position?: LogoPosition;- /** Multiplier supplied by a template. 1 preserves the original mark. */scale?: number;}> = ({ src, height, position = "top-left", scale = 1 }) => {if (!src) return null;@@ -33,13 +26,15 @@ export const Watermark: React.FC<{src={src.startsWith("http") ? src : staticFile(src)}style={{position: "absolute",- ...(position.startsWith("top-") ? { top: LOGO_INSET * s } : { bottom: LOGO_INSET * s }),+ ...(position.startsWith("top-")+ ? { top: LOGO_INSET * s }+ : { bottom: LOGO_INSET * s }),...(position.endsWith("-left")- ? { left: 108 * s }+ ? { left: LOGO_EDGE * s }: position.endsWith("-right")- ? { right: 108 * s }+ ? { right: LOGO_EDGE * s }: { left: "50%", transform: "translateX(-50%)" }),- width: 255 * s * scale,+ width: LOGO_WIDTH * s * scale,height: LOGO_HEIGHT * s * scale,objectFit: "contain",}}diff --git a/remotion/src/motion.ts b/remotion/src/motion.tsindex 2e627d2..cd63815 100644--- a/remotion/src/motion.ts+++ b/remotion/src/motion.ts@@ -37,6 +37,11 @@ export const MOTION: Record<string, Motion> = {branded: { enter: "none", exit: "none", duration: 0, feel: "linear" },/** Names arrive, hold, and get out of the way. */nameCard: { enter: "rise", exit: "fade", duration: 8, feel: "soft" },+ /**+ * A card takes the frame, so it has to arrive quickly enough not to read as+ * a slow wipe and leave without a gap where neither it nor the video is up.+ */+ card: { enter: "fade", exit: "fade", duration: 6, feel: "soft" },/** Always there, so it neither arrives nor leaves. */watermark: { enter: "none", exit: "none", duration: 0, feel: "linear" },};diff --git a/remotion/src/types.ts b/remotion/src/types.tsindex 629e380..4ca70b1 100644--- a/remotion/src/types.ts+++ b/remotion/src/types.ts@@ -4,6 +4,15 @@ export interface Word {end: number;confidence?: number;speaker?: string;+ /**+ * Worth reading even by somebody scrolling past with the sound off: a+ * number, a name, the word the sentence was built to land on.+ *+ * It rides on the word rather than arriving as a list of indices, because+ * every caption style chunks and splits the array before drawing it and an+ * index into the original would be wrong by then.+ */+ emphasis?: boolean;}export interface CaptionStyle {@@ -12,6 +21,11 @@ export interface CaptionStyle {fontFamily: string;color: string;activeColor: string;+ /**+ * What an emphasised word is coloured. Falls back to activeColor, so a+ * style that never sets one still marks emphasis rather than dropping it.+ */+ emphasisColor?: string;uppercase: boolean;wordsPerChunk: number;marginBottom: number;@@ -26,6 +40,19 @@ export type LogoPosition =| "bottom-center"| "bottom-right";+/**+ * The logo's box, in unscaled units.+ *+ * Watermark draws it and the branded captions keep clear of it, and those are+ * two files. Shared here so a logo that moves cannot leave the caption margin+ * guarding the place it used to be.+ */+export const LOGO_INSET = 180;+export const LOGO_EDGE = 108;+export const LOGO_WIDTH = 255;+export const LOGO_HEIGHT = 126;+export const LOGO_CAPTION_GAP = 24;+export interface CaptionProps {words: Word[];style: CaptionStyle;@@ -36,7 +63,18 @@ export interface CaptionProps {faceY?: number | null; // normalized 0-1 (0=top, 1=bottom)}-const FONT = "'DM Sans', sans-serif";+/**+ * The caption stack, in falling-back order.+ *+ * DM Sans draws Latin and nothing else, so the Noto families behind it are+ * what a Georgian or Russian show is actually rendered in. The browser picks+ * per glyph, so a Latin clip never leaves DM Sans.+ */+export const FONT = "'DM Sans', 'Noto Sans', 'Noto Sans Georgian', sans-serif";++/** A show's own font, ahead of the stack that covers what it cannot draw. */+export const fontStack = (family?: string | null): string =>+ family ? `'${family.replace(/'/g, "")}', ${FONT}` : FONT;// Caption geometry (font sizes, margins, insets) is authored for a 1920-tall// vertical canvas. Multiply pixel values by this factor so a shorter canvas@@ -52,6 +90,7 @@ export const STYLES: Record<string, CaptionStyle> = {fontFamily: FONT,color: "#FFFFFF",activeColor: "#FFFF00",+ emphasisColor: "#3B9CFF",uppercase: true,wordsPerChunk: 3,marginBottom: 400,@@ -62,6 +101,7 @@ export const STYLES: Record<string, CaptionStyle> = {fontFamily: FONT,color: "rgba(255,255,255,0.4)",activeColor: "#FFFFFF",+ emphasisColor: "#3B9CFF",uppercase: false,wordsPerChunk: 5,marginBottom: 400,@@ -72,6 +112,7 @@ export const STYLES: Record<string, CaptionStyle> = {fontFamily: FONT,color: "#FFFFFF",activeColor: "#FFFFFF",+ emphasisColor: "#7FD1FF",uppercase: false,wordsPerChunk: 6,marginBottom: 200,@@ -82,6 +123,7 @@ export const STYLES: Record<string, CaptionStyle> = {fontFamily: FONT,color: "#FFFFFF",activeColor: "#FFFFFF",+ emphasisColor: "#3B9CFF",uppercase: false,wordsPerChunk: 3,marginBottom: 420,