AI Content Repurposing API: Summaries, Threads, Blog Posts

Turn any transcript into summaries, chapter markers, quote collections, Twitter threads, blog posts, or show notes — all via a single endpoint with a mode selector.

The repurpose endpoint takes a completed transcript and generates derivative content in a single API call. Results are returned synchronously — average latency is 3–8 seconds depending on transcript length and mode.

Endpoint

bash
POST /api/v1/transcriptions/{id}/repurpose Scope: repurpose:run

Modes

ModeWhat you getTypical length
summaryParagraph summary of the transcript3–5 sentences
chaptersTimestamped chapter markers with titles5–15 entries
quotesTop 5–10 most quotable lines pulled verbatim5–10 quotes
show-notesPodcast show notes with intro, key topics, links placeholder200–400 words
threadTwitter/X thread with hooks and numbered tweets8–15 tweets
blogFull blog post with sections and a CTA600–1200 words

Example — generate a Twitter thread

const result = await client.transcriptions.repurpose(99, 'thread'); console.log(result.content);
json
{ "success": true, "data": { "mode": "thread", "content": "1/ Never gonna give you up — breaking down the most misunderstood song of the 80s.\n\n2/ Here's what Rick actually meant when he wrote it...", "generatedAt": "2026-06-17T10:01:00.000Z" } }

Example — chapter markers

const chapters = await client.transcriptions.chapters(99); console.log(chapters.chapters); // [{ title: 'Intro', startSeconds: 0 }, { title: 'The hook', startSeconds: 82 }]
json
{ "success": true, "data": { "mode": "chapters", "content": "00:00 Intro\n01:22 The hook\n03:45 Verse breakdown\n...", "chapters": [ { "title": "Intro", "startSeconds": 0 }, { "title": "The hook", "startSeconds": 82 } ], "generatedAt": "2026-06-17T10:01:00.000Z" } }

SDK convenience aliases

// Generic const result = await client.transcriptions.repurpose(99, 'thread'); // Convenience aliases const summary = await client.transcriptions.summary(99); const chapters = await client.transcriptions.chapters(99); const quotes = await client.transcriptions.quotes(99);

Notes

  • Repurpose calls require the transcript to be in completed status. A 422 is returned otherwise.
  • Results are not stored — generate on demand and cache on your end if needed.
  • Average latency: ~3–8s depending on transcript length and mode.

FAQ

What repurpose modes are available?

Six modes are supported: summary, chapters, quotes, show-notes, thread, and blog. See the modes table above for output format and typical length.

Can I use my own prompt or tone instructions?

Custom prompt support is on the roadmap. For now, each mode uses a purpose-built prompt optimized for that content type.

How long does repurposing take?

Most modes return in 3–8 seconds for a typical podcast episode (30–60 minutes). Longer transcripts or the blog mode (which produces 600–1200 words) may take up to 15 seconds.