CallMissed
Back to all models
Text to Speechbudgetopen-source

MeloTTS

by MyShell AI · Released 2024

Lightweight open-source TTS from MyShell AI. English + French. Best for short utterances and high-throughput notification / transactional speech.

Text to Speech

MeloTTS

Powered by MyShell AI · Open-source multilingual TTS

Context Window

N/A

Parameters

Open-source (~6M params per language)

Max Output

N/A

Category

Text to Speech

Overview

MeloTTS is MyShell AI's open-source multilingual TTS model, optimized for fast generation of short utterances. It supports English and French and is best suited for high-throughput, short-text use cases like notification audio, transactional confirmations, and lightweight voice prompts.

Deployed via Cloudflare Workers AI, it returns base64-encoded MP3 audio in a single response. Latency is very low for short inputs (<5 seconds), and the per-minute cost is the cheapest TTS available on the platform.

At $0.05 per 10K characters, it is roughly 8× cheaper than Aura 2 and 10× cheaper than Bulbul. Pick MeloTTS when cost matters more than voice variety or prosody control — for example, sending many short notification clips at scale.

Pricing

MetricPrice
Price /10K chars₹5.0000

1 credit = ₹1 = $0.01 USD. Prices shown from provider; CallMissed passes through with ~35% markup.

Key Highlights

  • Cheapest TTS on the platform
  • English + French support
  • Optimized for short utterances
  • Open-source weights (MIT-licensed)

Benchmarks

BenchmarkScore
Languages2
Cost$0.05
Best for<5s

Technical Details

  • Runs on Cloudflare Workers AI (`@cf/myshell-ai/melotts`)
  • Returns base64-encoded MP3 in JSON response
  • Body: {"prompt": "<text>", "lang": "en"|"fr"}
  • Best suited for short utterances; longer inputs may degrade in quality

Strengths

  • Cheapest TTS available
  • Fast generation for short utterances
  • Open-source — well understood, no vendor lock-in

Limitations

  • Only English + French
  • No voice variety — single voice per language
  • Quality degrades on long inputs (>30s)
  • No prosody / SSML control

Use Cases

Notification audioTransactional confirmationsLightweight voice promptsHigh-volume short-text TTS

API Example

curl https://api.callmissed.com/v1/audio/speech \
  -H "Authorization: Bearer cm_YOUR_KEY" \
  -d '{"model": "melotts", "input": "Your order has been shipped.", "language": "en"}' \
  --output speech.mp3

Endpoint: POST /v1/audio/speech · Model ID: melotts

Try MeloTTS now

Get 1000 free API credits on signup. No credit card required.

Start free Read docs