From fe761ce0f05b6cc166f37b30b5def8c4218ab2a7 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Mon, 14 Sep 2026 22:14:28 +0000
Subject: [PATCH] Add toggle to include or omit speaker names on transcript
 export

SRT/VTT/TXT/MD/DOCX/PDF previously always prefixed cues or turns with
speaker labels. Make that optional (default on); JSON still keeps speaker
ids and names for round-trip re-import.

Co-authored-by: Wassim Gharbi <wassgha@gmail.com>
---
 components/ExportDialog.tsx        | 17 ++++++++
 lib/i18n/messages/de.ts            |  1 +
 lib/i18n/messages/en.ts            |  1 +
 lib/i18n/messages/es.ts            |  1 +
 lib/i18n/messages/fr.ts            |  1 +
 lib/i18n/messages/ja.ts            |  1 +
 lib/i18n/messages/ko.ts            |  1 +
 lib/i18n/messages/pt.ts            |  1 +
 lib/i18n/messages/zh-CN.ts         |  1 +
 lib/i18n/messages/zh-TW.ts         |  1 +
 lib/serializeTranscript.ts         | 70 +++++++++++++++++++++++-------
 tests/serialize-transcript-test.ts | 41 +++++++++++++++++
 12 files changed, 121 insertions(+), 16 deletions(-)

diff --git a/components/ExportDialog.tsx b/components/ExportDialog.tsx
index 7023f95..abe97bc 100644
--- a/components/ExportDialog.tsx
+++ b/components/ExportDialog.tsx
@@ -113,6 +113,7 @@ export default function ExportDialog() {
   const [textFormat, setTextFormat] = useState<TranscriptFormat>("txt");
   const [includeTimestamps, setIncludeTimestamps] = useState(false);
   const [shortCues, setShortCues] = useState(true);
+  const [speakerLabels, setSpeakerLabels] = useState(true);
   const [timelineFormat, setTimelineFormat] =
     useState<TimelineExportFormat>("resolve");
   const [timelineFrameRate, setTimelineFrameRate] =
@@ -273,6 +274,7 @@ export default function ExportDialog() {
 
   const textSupportsTimestamps = DOC_FORMATS.has(textFormat);
   const textSupportsShortCues = SUBTITLE_FORMATS.has(textFormat);
+  const textSupportsSpeakerLabels = textFormat !== "json";
 
   const exportText = useCallback(() => {
     // Read live store state at click time so a rename / replace-in-project
@@ -287,6 +289,7 @@ export default function ExportDialog() {
         speakers: speakersFromWords(s.words, s.speakers),
         ...(textSupportsTimestamps ? { timestamps: includeTimestamps } : {}),
         ...(textSupportsShortCues ? { shortCues } : {}),
+        ...(textSupportsSpeakerLabels ? { speakerLabels } : {}),
       });
       setError(null);
       trackEvent("export_completed", {
@@ -294,6 +297,7 @@ export default function ExportDialog() {
         format: textFormat,
         ...(textSupportsTimestamps ? { timestamps: includeTimestamps } : {}),
         ...(textSupportsShortCues ? { shortCues } : {}),
+        ...(textSupportsSpeakerLabels ? { speakerLabels } : {}),
       });
     } catch (err) {
       setError(err instanceof Error ? err.message : en["error.export"]);
@@ -302,8 +306,10 @@ export default function ExportDialog() {
     textFormat,
     textSupportsTimestamps,
     textSupportsShortCues,
+    textSupportsSpeakerLabels,
     includeTimestamps,
     shortCues,
+    speakerLabels,
     baseName,
   ]);
 
@@ -514,6 +520,17 @@ export default function ExportDialog() {
                 <span>{t("export.includeTimestamps")}</span>
               </label>
             )}
+            {textSupportsSpeakerLabels && (
+              <label className="flex cursor-pointer items-center gap-2.5 rounded-lg bg-zinc-50 px-3 py-2.5 text-sm text-zinc-700 dark:bg-zinc-800/60 dark:text-zinc-200">
+                <input
+                  type="checkbox"
+                  checked={speakerLabels}
+                  onChange={(e) => setSpeakerLabels(e.target.checked)}
+                  className="h-3.5 w-3.5 rounded border-zinc-300 text-zinc-900 focus:ring-zinc-400 dark:border-zinc-600 dark:bg-zinc-900"
+                />
+                <span>{t("export.speakerLabels")}</span>
+              </label>
+            )}
             {textSupportsShortCues && (
               <label className="flex cursor-pointer items-center gap-2.5 rounded-lg bg-zinc-50 px-3 py-2.5 text-sm text-zinc-700 dark:bg-zinc-800/60 dark:text-zinc-200">
                 <input
diff --git a/lib/i18n/messages/de.ts b/lib/i18n/messages/de.ts
index 593c733..2e5bee4 100644
--- a/lib/i18n/messages/de.ts
+++ b/lib/i18n/messages/de.ts
@@ -140,6 +140,7 @@ export const de: Record<MessageKey, string> = {
   "export.statCuts": "Schnitte",
   "export.statEdited": "Bearbeitet",
   "export.includeTimestamps": "Zeitstempel einbeziehen",
+  "export.speakerLabels": "Sprechernamen einbeziehen",
   "export.shortCues": "Kurze Untertitel (≤5 s)",
   "export.transcriptHelp": "Text mit Sprecherlabels und entfernten Schnitten.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/en.ts b/lib/i18n/messages/en.ts
index 582a74f..9920ea2 100644
--- a/lib/i18n/messages/en.ts
+++ b/lib/i18n/messages/en.ts
@@ -141,6 +141,7 @@ export const en = {
   "export.statCuts": "Cuts",
   "export.statEdited": "Edited",
   "export.includeTimestamps": "Include timestamps",
+  "export.speakerLabels": "Include speaker names",
   "export.shortCues": "Short cues (≤5s)",
   "export.transcriptHelp": "Speaker-labeled text with cuts removed.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/es.ts b/lib/i18n/messages/es.ts
index bedb5fa..d0f5482 100644
--- a/lib/i18n/messages/es.ts
+++ b/lib/i18n/messages/es.ts
@@ -140,6 +140,7 @@ export const es: Record<MessageKey, string> = {
   "export.statCuts": "Cortes",
   "export.statEdited": "Editado",
   "export.includeTimestamps": "Incluir marcas de tiempo",
+  "export.speakerLabels": "Incluir nombres de hablantes",
   "export.shortCues": "Subtítulos cortos (≤5 s)",
   "export.transcriptHelp": "Texto con etiquetas de hablante y cortes eliminados.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/fr.ts b/lib/i18n/messages/fr.ts
index 40e25f5..28fcd31 100644
--- a/lib/i18n/messages/fr.ts
+++ b/lib/i18n/messages/fr.ts
@@ -140,6 +140,7 @@ export const fr: Record<MessageKey, string> = {
   "export.statCuts": "Coupes",
   "export.statEdited": "Monté",
   "export.includeTimestamps": "Inclure les horodatages",
+  "export.speakerLabels": "Inclure les noms des locuteurs",
   "export.shortCues": "Sous-titres courts (≤5 s)",
   "export.transcriptHelp": "Texte avec étiquettes de locuteur et coupes retirées.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/ja.ts b/lib/i18n/messages/ja.ts
index 487ae20..583cf1b 100644
--- a/lib/i18n/messages/ja.ts
+++ b/lib/i18n/messages/ja.ts
@@ -140,6 +140,7 @@ export const ja: Record<MessageKey, string> = {
   "export.statCuts": "カット",
   "export.statEdited": "編集後",
   "export.includeTimestamps": "タイムスタンプを含める",
+  "export.speakerLabels": "話者名を含める",
   "export.shortCues": "短い字幕（≤5秒）",
   "export.transcriptHelp": "話者ラベル付きテキスト。カット部分は除外されます。",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/ko.ts b/lib/i18n/messages/ko.ts
index 5fae2fe..cb19160 100644
--- a/lib/i18n/messages/ko.ts
+++ b/lib/i18n/messages/ko.ts
@@ -140,6 +140,7 @@ export const ko: Record<MessageKey, string> = {
   "export.statCuts": "컷",
   "export.statEdited": "편집본",
   "export.includeTimestamps": "타임스탬프 포함",
+  "export.speakerLabels": "화자 이름 포함",
   "export.shortCues": "짧은 자막 (≤5초)",
   "export.transcriptHelp": "화자 라벨이 포함된 텍스트이며 컷한 내용은 제거됩니다.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/pt.ts b/lib/i18n/messages/pt.ts
index 1fa457a..3e2056d 100644
--- a/lib/i18n/messages/pt.ts
+++ b/lib/i18n/messages/pt.ts
@@ -140,6 +140,7 @@ export const pt: Record<MessageKey, string> = {
   "export.statCuts": "Cortes",
   "export.statEdited": "Editado",
   "export.includeTimestamps": "Incluir marcas de tempo",
+  "export.speakerLabels": "Incluir nomes dos falantes",
   "export.shortCues": "Legendas curtas (≤5 s)",
   "export.transcriptHelp": "Texto com rótulos de falante e cortes removidos.",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/zh-CN.ts b/lib/i18n/messages/zh-CN.ts
index 39344b2..78a7ee7 100644
--- a/lib/i18n/messages/zh-CN.ts
+++ b/lib/i18n/messages/zh-CN.ts
@@ -142,6 +142,7 @@ export const zhCN: Record<MessageKey, string> = {
   "export.statCuts": "剪除",
   "export.statEdited": "成片",
   "export.includeTimestamps": "包含时间戳",
+  "export.speakerLabels": "包含说话人姓名",
   "export.shortCues": "短字幕（≤5 秒）",
   "export.transcriptHelp": "带说话人标签的文本，已移除剪除内容。",
   "export.transcriptHelpTimestamps":
diff --git a/lib/i18n/messages/zh-TW.ts b/lib/i18n/messages/zh-TW.ts
index c94fa01..92ae4b5 100644
--- a/lib/i18n/messages/zh-TW.ts
+++ b/lib/i18n/messages/zh-TW.ts
@@ -140,6 +140,7 @@ export const zhTW: Record<MessageKey, string> = {
   "export.statCuts": "剪除",
   "export.statEdited": "成片",
   "export.includeTimestamps": "包含時間戳",
+  "export.speakerLabels": "包含說話人姓名",
   "export.shortCues": "短字幕（≤5 秒）",
   "export.transcriptHelp": "帶有說話人標籤的文字，已移除剪除內容。",
   "export.transcriptHelpTimestamps":
diff --git a/lib/serializeTranscript.ts b/lib/serializeTranscript.ts
index c65b241..a951f3a 100644
--- a/lib/serializeTranscript.ts
+++ b/lib/serializeTranscript.ts
@@ -37,6 +37,12 @@ export interface SerializeOptions {
    * on speaker changes and pause gaps (legacy long cues).
    */
   shortCues?: boolean;
+  /**
+   * When true (default for SRT/VTT/TXT/MD/DOCX/PDF), prefix cues / turns with
+   * speaker names (`Speaker 1: …`, VTT `<v>`, document headings). When false,
+   * export dialogue only. JSON always keeps speaker ids/names for re-import.
+   */
+  speakerLabels?: boolean;
 }
 
 interface Cue {
@@ -108,10 +114,11 @@ export function serializeTranscript(
     throw new Error(en["error.noWords"]);
   }
   const shortCues = options.shortCues !== false;
+  const speakerLabels = options.speakerLabels !== false;
   const cues = wordsToCues(prepared, shortCues);
   return format === "vtt"
-    ? serializeVtt(cues, speakers, shortCues)
-    : serializeSrt(cues, speakers, shortCues);
+    ? serializeVtt(cues, speakers, shortCues, speakerLabels)
+    : serializeSrt(cues, speakers, shortCues, speakerLabels);
 }
 
 /** Serialize to DOCX or PDF (binary). */
@@ -122,8 +129,11 @@ export function serializeTranscriptBinary(
 ): Uint8Array {
   const speakers = speakersFromWords(words, options.speakers ?? []);
   const turns = buildDocumentTurns(words, { ...options, speakers });
-  if (format === "docx") return serializeDocx(turns, speakers, options.timestamps);
-  return serializePdf(turns, speakers, options.timestamps);
+  const speakerLabels = options.speakerLabels !== false;
+  if (format === "docx") {
+    return serializeDocx(turns, speakers, options.timestamps, speakerLabels);
+  }
+  return serializePdf(turns, speakers, options.timestamps, speakerLabels);
 }
 
 /** Trigger a browser download of the serialized transcript / subtitles. */
@@ -198,11 +208,17 @@ function serializeDocument(
   const speakers = speakersFromWords(words, options.speakers ?? []);
   const turns = buildDocumentTurns(words, { ...options, speakers });
   const withTs = Boolean(options.timestamps);
+  const withLabels = options.speakerLabels !== false;
 
   if (format === "txt") {
     return (
       turns
         .map((turn) => {
+          if (!withLabels) {
+            return withTs
+              ? `[${formatTranscriptTimestamp(turn.start)}] ${turn.text}`
+              : turn.text;
+          }
           const label = speakerLabel(speakers, turn.speaker);
           const prefix = withTs
             ? `[${formatTranscriptTimestamp(turn.start)}] ${label}`
@@ -216,6 +232,11 @@ function serializeDocument(
   return (
     turns
       .map((turn) => {
+        if (!withLabels) {
+          return withTs
+            ? `**[${formatTranscriptTimestamp(turn.start)}]**\n\n${turn.text}`
+            : turn.text;
+        }
         const label = speakerLabel(speakers, turn.speaker);
         const heading = withTs
           ? `**[${formatTranscriptTimestamp(turn.start)}] ${label}**`
@@ -237,10 +258,20 @@ function escapeXml(text: string): string {
 function serializeDocx(
   turns: DocumentTurn[],
   speakers: SpeakerInfo[],
-  timestamps?: boolean
+  timestamps?: boolean,
+  speakerLabels = true
 ): Uint8Array {
   const paragraphs = turns
     .map((turn) => {
+      const body = `<w:p><w:pPr><w:spacing w:after="240"/></w:pPr><w:r><w:t xml:space="preserve">${escapeXml(turn.text)}</w:t></w:r></w:p>`;
+      if (!speakerLabels) {
+        if (!timestamps) return body;
+        const heading = `[${formatTranscriptTimestamp(turn.start)}]`;
+        return [
+          `<w:p><w:pPr><w:spacing w:after="80"/></w:pPr><w:r><w:rPr><w:b/></w:rPr><w:t xml:space="preserve">${escapeXml(heading)}</w:t></w:r></w:p>`,
+          body,
+        ].join("");
+      }
       const label = speakerLabel(speakers, turn.speaker);
       const heading = timestamps
         ? `[${formatTranscriptTimestamp(turn.start)}] ${label}`
@@ -249,7 +280,7 @@ function serializeDocx(
         // Speaker heading (bold)
         `<w:p><w:pPr><w:spacing w:after="80"/></w:pPr><w:r><w:rPr><w:b/></w:rPr><w:t xml:space="preserve">${escapeXml(heading)}</w:t></w:r></w:p>`,
         // Body
-        `<w:p><w:pPr><w:spacing w:after="240"/></w:pPr><w:r><w:t xml:space="preserve">${escapeXml(turn.text)}</w:t></w:r></w:p>`,
+        body,
       ].join("");
     })
     .join("");
@@ -447,7 +478,8 @@ function wrapCueLines(text: string, maxChars: number): string[] {
 function serializeSrt(
   cues: Cue[],
   speakers: SpeakerInfo[],
-  shortCues: boolean
+  shortCues: boolean,
+  speakerLabels: boolean
 ): string {
   return (
     cues
@@ -456,7 +488,7 @@ function serializeSrt(
           String(i + 1),
           `${formatSrtTimestamp(cue.start)} --> ${formatSrtTimestamp(cue.end)}`,
         ];
-        if (cue.speaker >= 0) {
+        if (speakerLabels && cue.speaker >= 0) {
           lines.push(
             formatCueBody(
               cue.text,
@@ -476,7 +508,8 @@ function serializeSrt(
 function serializeVtt(
   cues: Cue[],
   speakers: SpeakerInfo[],
-  shortCues: boolean
+  shortCues: boolean,
+  speakerLabels: boolean
 ): string {
   const body = cues
     .map((cue) => {
@@ -485,7 +518,7 @@ function serializeVtt(
         ? wrapCueLines(cue.text, CUE_LINE_CHARS).join("\n")
         : cue.text;
       const text =
-        cue.speaker >= 0
+        speakerLabels && cue.speaker >= 0
           ? `<v ${speakerLabel(speakers, cue.speaker)}>${dialogue}`
           : dialogue;
       return `${timing}\n${text}`;
@@ -604,7 +637,8 @@ function wrapPdfLine(text: string, maxChars: number): string[] {
 function serializePdf(
   turns: DocumentTurn[],
   speakers: SpeakerInfo[],
-  timestamps?: boolean
+  timestamps?: boolean,
+  speakerLabels = true
 ): Uint8Array {
   const pageWidth = 612;
   const pageHeight = 792;
@@ -617,11 +651,15 @@ function serializePdf(
   type PdfLine = { text: string; bold: boolean };
   const allLines: PdfLine[] = [];
   for (const turn of turns) {
-    const label = speakerLabel(speakers, turn.speaker);
-    const heading = timestamps
-      ? `[${formatTranscriptTimestamp(turn.start)}] ${label}`
-      : label;
-    allLines.push({ text: heading, bold: true });
+    if (speakerLabels || timestamps) {
+      const label = speakerLabel(speakers, turn.speaker);
+      const heading = !speakerLabels
+        ? `[${formatTranscriptTimestamp(turn.start)}]`
+        : timestamps
+          ? `[${formatTranscriptTimestamp(turn.start)}] ${label}`
+          : label;
+      allLines.push({ text: heading, bold: true });
+    }
     for (const line of wrapPdfLine(turn.text, maxChars)) {
       allLines.push({ text: line, bold: false });
     }
diff --git a/tests/serialize-transcript-test.ts b/tests/serialize-transcript-test.ts
index ed73b27..8a0028a 100644
--- a/tests/serialize-transcript-test.ts
+++ b/tests/serialize-transcript-test.ts
@@ -435,4 +435,45 @@ const sample: Word[] = [
   console.log("max cue chars: ok");
 }
 
+{
+  // speakerLabels: false omits names from captions and documents.
+  const srt = serializeTranscript(sample, "srt", {
+    duration: 10,
+    speakerLabels: false,
+  });
+  assert(srt.includes("Hello world"), `srt text\n${srt}`);
+  assert(srt.includes("How are you"), `srt text 2\n${srt}`);
+  assert(!/Speaker \d+:/.test(srt), `srt should omit speaker prefix\n${srt}`);
+
+  const vtt = serializeTranscript(sample, "vtt", {
+    duration: 10,
+    speakerLabels: false,
+  });
+  assert(vtt.includes("Hello world"), `vtt text\n${vtt}`);
+  assert(!vtt.includes("<v "), `vtt should omit voice tags\n${vtt}`);
+
+  const txt = serializeTranscript(sample, "txt", {
+    duration: 10,
+    speakerLabels: false,
+  });
+  assert(txt.includes("Hello world"), `txt text\n${txt}`);
+  assert(!/Speaker \d+:/.test(txt), `txt should omit labels\n${txt}`);
+
+  const md = serializeTranscript(sample, "md", {
+    duration: 10,
+    speakerLabels: false,
+    timestamps: true,
+  });
+  assert(md.includes("Hello world"), `md text\n${md}`);
+  assert(md.includes("**[0:01]**"), `md keeps timestamps without names\n${md}`);
+  assert(!md.includes("Speaker"), `md should omit speaker names\n${md}`);
+
+  // JSON always keeps speakers for re-import regardless of the toggle.
+  const json = serializeTranscript(sample, "json", { speakerLabels: false });
+  const data = JSON.parse(json);
+  assert(data.speakers?.[0]?.name === "Speaker 1", "json keeps speaker names");
+  assert(data.words[0].speaker === 0, "json keeps speaker ids");
+  console.log("speakerLabels off: ok");
+}
+
 console.log("ALL SERIALIZE TRANSCRIPT TESTS PASSED");