Provider pricing comparison
The extension is free for individuals. You only pay your chosen cloud provider for the characters you synthesize (every character, including spaces, counts). Here's how the four cloud providers compare, alongside the OpenAI-compatible option for local or proxied servers. Prices are approximate, in USD, and change over time; always confirm on the official pricing pages linked below.
Free tier: 5M standard + 1M neural chars/month, first 12 months only
Cost per 1M characters
- Standard: $4
- Neural: $16
- Generative: $30
- Long-form: $100
Free tier: 0.5M neural chars/month on the Free (F0) tier; renews every month
Cost per 1M characters
- Neural: ~$15-16
- HD / custom voices: higher
Free tier: 4M standard + 1M WaveNet chars/month; renews every month (Studio-class voices have a much smaller separate free allowance)
Cost per 1M characters
- Standard: ~$4
- WaveNet / Neural2: ~$16
- Chirp HD: higher
- Chirp 3 HD / Gemini-TTS: ~$30
- Studio: ~$160
Free tier: No free tier; pay as you go
Cost per 1M characters
- tts-1: ~$15
- tts-1-hd: ~$30
- gpt-4o-mini-tts: token-billed; ~$12/1M audio tokens (about the same as tts-1 in practice)
Free tier: Local engines are free to run; gateways bill per their backend
Cost per 1M characters
- Local (LocalAI, Speaches): free
- Via a gateway (LiteLLM, Groq): backend rates
What does that mean in practice?
One million characters is roughly 10-15 hours of listening, about two average novels. At $4 per million (standard voices), reading a typical 1,500-word article costs about 4 hundredths of a cent; at $16 per million (neural), about a sixth of a cent. Casual daily use usually stays within the Azure or Google monthly free tiers entirely.
Cost-optimization tips
- Use standard voices for most content. They cost a quarter of neural voices and are fine for skimming articles.
- Reserve neural voices for important audio: long reads, downloads you will keep, or content you listen to closely.
- Use generative, HD, and long-form voices sparingly. They cost 2-25x more than standard voices.
- Ride the monthly free tiers. Azure (0.5M neural) and Google (1M WaveNet + 4M standard) reset every month. Combine providers and casual use can be free forever.
- Monitor usage in your provider's console (AWS Console → Polly usage reports, Azure Portal metrics, Google Cloud billing reports, OpenAI usage dashboard).
- Set up billing alerts so you're notified before charges grow. Every provider supports budgets or spending alerts.
Ready to connect a provider? Start with a setup guide: Amazon Polly, Azure Speech, Google Cloud TTS, OpenAI, or an OpenAI-compatible server.