⚡ TL;DR GPT Transcribe (default) handles names and jargon with hints, up to 25 MB. Gemini 3.5 Transcribe auto-detects 85+ languages, up to 14 MB. Split long recordings into segments.
Typing is slow. Talking is not.
📝 Uses
Meeting notes and action items, dictation, searchable voice memos, captions and spoken requests.
🧠 Models
GPT Transcribe is accurate with specialist vocabulary, takes context hints and accepts files up to 25 MB. Gemini 3.5 Transcribe covers 85+ languages, handles filler words and accepts up to 14 MB.
✂️ Long Recordings
Ask Lovable to split audio into segments to stay under the limits.
🔀 Need a Conversation?
For two-way voice, use live voice instead.
💡 Pro tip Pass product names as keywords to GPT Transcribe for better accuracy.
Want a meeting assistant built? Lord of the Wix can help. ✨
❓ Lovable speech to text FAQ
What is the default model?
GPT Transcribe.
How many languages does Gemini detect?
85+.
What is the GPT Transcribe file limit?
25 MB.
Do transcripts stream?
Yes.
What about long recordings?
Split them into segments.
🙏 Credit: Based on public information from the Lovable documentation. Lovable is a product of Lovable. Source: docs.lovable.dev. Lord of the Wix is an independent agency and is not affiliated with Lovable.