Automatic splitting or your own sentences
Send text and let Kolimina split it, or send sentences[] to keep your own editing units.
text is split into sentences exactly as on the website. sentences[] keeps your boundaries (and your client_sentence_id for each), so you can correct the same unit later. Both are priced the same: sentences[] costs what the same text typed one sentence per line costs.
- Up to 5,000 spoken characters per narration (8,000 including expressions) and 200 sentences.
- Each sentence: up to 300 spoken characters; a word longer than 40 letters is refused.
- Exactly one of text or sentences.
curl -s -X POST "https://kolimina.com/api/public/v1/narrations" \
-H "Authorization: Bearer $KOLIMINA_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: units-1" \
-d '{"voice_id": "voice_…", "sentences": [{"client_sentence_id": "s1", "text": "פתיחה."}, {"client_sentence_id": "s2", "text": "המשך."}]}'Preview before you pay
POST /narration-previews takes the same body and runs the text through every correction stage - no audio and no narration credits. Poll it, show the processed sentences and the estimated credits, then create exactly that snapshot with {"preview_id": …}. A preview lives 30 minutes and creates one narration; if it expired, was used, or the voice, style or pricing changed, creation answers 409 preview_expired_or_stale.
curl -s -X POST "https://kolimina.com/api/public/v1/narration-previews" \
-H "Authorization: Bearer $KOLIMINA_API_KEY" \
-H "Content-Type: application/json" \
-d '{"voice_id": "voice_…", "text": "שלום עולם."}'curl -s -X POST "https://kolimina.com/api/public/v1/narrations" \
-H "Authorization: Bearer $KOLIMINA_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: from-preview-1" \
-d '{"preview_id": "prev_…"}'