blink-new Blink Speech (Text-to-Speech)
Generate natural-sounding speech from text using Blink AI (OpenAI TTS). No API key needed — charged to your workspace credits.
Generate speech and save to file
blink ai speech "Hello! This is your AI agent speaking." --voice alloy --output /data/speech.mp3
Choose a different voice
blink ai speech "Good morning! Here's your daily briefing." --voice nova --output /data/briefing.mp3
Available voices
| Voice | Character |
|---|---|
alloy |
Neutral, balanced (default) |
echo |
Male, clear |
fable |
Storytelling, expressive |
onyx |
Deep, authoritative |
nova |
Warm, female |
shimmer |
Soft, female |
Read a text file aloud
TEXT=$(cat /data/report.txt | head -c 4000)
blink ai speech "$TEXT" --voice alloy --output /data/report_audio.mp3
echo "Saved to /data/report_audio.mp3"
Common use cases
- "Read this summary aloud" → generate speech, save file
- "Create a voice message saying X" → speak text, save mp3
- "Narrate this blog post" → read text, generate audio file
- "Make my agent greet users with voice" → generate welcome audio
Notes
- Max input: 4096 characters per call (split longer text)
- Output is saved as mp3 by default