feature

Speech to Text now supports natural-language transcript editing

Batch and realtime transcription endpoints accept edit instructions up to 2,000 characters, returning results in edited_transcript. Developers can now programmatically refine transcriptions without re-running models.

Stacks this affects

ElevenLabs ships in 5 tested stacks here.

Faceless YouTube
Video · Creator
$92/mo
Faceless YouTube (budget)
Video · Creator
$6/mo
TikTok character
Video · Creator
$92/mo
TikTok character (audio + text)
Video · Creator
$42/mo
UGC ad creative
Marketing · Ads
$67/mo
All ElevenLabs changes and pricing

More ElevenLabs

releaseEleven v4 and Eleven v4 Turbo speech models released with improved voice cloning and real-time supportshutdownSora 2 and Sora 2 Pro APIs discontinued by OpenAIshutdownSeedance 1.5 Pro model deprecated by ByteDanceshutdownOpenAI discontinues Sora 2 and Sora 2 Pro models, retiring the Sora API on September 24, 2026shutdownByteDance retires Seedance 1.5 Pro on November 11, 2026