TTS Script Calculator
Paste your script. Compare 8 providers. See the real cost.
Cost comparison
Estimated Costs
ElevenLabs Starter
ElevenLabs Creator
* Calculates raw character strings including spaces. Subscription costs show base quota utilization; overage rates vary by provider.
Why does this calculator exist?
Because comparing text-to-speech API costs across 8 providers and 11 pricing tiers is a nightmare. Some charge per character, some per audio minute, some bundle characters into monthly subscription quotas. Amazon Polly, Azure, and Google publish per-million-character rates. ElevenLabs sells monthly plans with soft caps. Cartesia charges per character but markets on latency. Nobody makes it easy to compare.
This tool normalises every provider to the same metric: your exact character count × their published rate. Paste your script, get the real number. No signup, no paywall.
Pay-as-you-go vs. subscription
OpenAI, Google, Amazon Polly, Azure, Deepgram, and Cartesia all charge per character with no commitment. ElevenLabs uses monthly subscription quotas — $6/mo for 30k characters up to $330/mo for 500k. If you exceed your plan, overage runs $0.30 per 1,000 characters ($300/1M). For low-volume creators, pay-as-you-go is almost always cheaper. For predictable production pipelines, subscriptions can work — but only if you consistently use the full quota.
How much does a 10-minute video cost?
At 130 words per minute, a 10-minute script is roughly 1,300 words or 7,500 characters. On Amazon Polly Standard or Google Standard, that costs about 3 cents. On OpenAI tts-1, roughly 11 cents. On Cartesia Sonic, 38 cents. On ElevenLabs Creator, it burns 7.5% of your entire monthly quota. The spread from cheapest to most expensive is over 50×.
Which TTS API has the best quality?
ElevenLabs remains the quality benchmark for expressiveness and voice cloning. Cartesia Sonic leads on latency (sub-100ms time-to-first-audio) for real-time voice agents. Azure Neural HD and Google Chirp 3 HD offer premium quality at cloud-provider scale. OpenAI tts-1 is the most widely integrated option with solid, reliable output. Amazon Polly Generative uses an LLM-based approach that is competitive for English narration.
Free tiers worth knowing about
Azure offers 500,000 characters/month free on Neural Standard voices — no expiration. Google gives 1M characters/month free on WaveNet and Neural2. Amazon Polly includes 5M Standard characters/month for the first year. Deepgram provides $200 in signup credits. ElevenLabs' free plan is limited to 10,000 characters/month — roughly 7 minutes of audio.