Switch from OpenAI text to speech
Cost next to OpenAI, the lines to switch, and the limits.
Prices read 2026-10-08 from the public pricing page of OpenAI, per million characters. We have not measured OpenAI speed or quality.
Cost per million characters
| Plan or model | Per million characters | 10 million characters | Times our $3 |
|---|---|---|---|
| tts-1 | $15 | $150 | 5x |
| tts-1-hd | $30 | $300 | 10x |
| Speakvora catalog voices | $3 | $30 | 1x |
Switch in a few lines
The OpenAI client works when you set the base URL to https://speakvora.com/api/v1 and use your Speakvora key; OpenAI voice names map to the closest catalog voice.
from openai import OpenAI
client = OpenAI(base_url="https://speakvora.com/api/v1", api_key="YOUR_SPEAKVORA_KEY")
client.audio.speech.create(model="tts-1", voice="alloy", input="Your order is ready.")What is different
- Plain text only: no SSML, speech marks or lexicons.
- mp3 or AAC audio on the compatible endpoints (the content type says which).
- Fewer voices and languages.
- New accounts get 100,000 characters free; billing is by the character with spending caps.
Speed we did measure: text to speech latency. Full guide: switch from Amazon Polly, OpenAI or ElevenLabs.
Questions
Can I use Speakvora instead of OpenAI text to speech?
For plain text to speech, yes. The OpenAI client works when you set the base URL to https://speakvora.com/api/v1 and use your Speakvora key; OpenAI voice names map to the closest catalog voice. Features Speakvora does not have are listed on this page and return an error.
How much cheaper is Speakvora than OpenAI?
Speakvora catalog voices are $3 per million characters. OpenAI lists tts-1 at $15 and tts-1-hd at $30 per million characters. Prices were read on 2026-10-08 from the provider's pricing page; plans, credits and volume discounts change what you actually pay, so check yours.
Is Speakvora faster than OpenAI?
We have not measured OpenAI, so we make no speed claim against it. Our measured time to first audio and the comparison we did run (Amazon Polly) are on the latency page.
Do I lose anything by switching?
Yes: fewer voices and languages, no SSML or speech marks, mp3 or AAC output for the compatible endpoints, and no uptime guarantee. Listen to samples first and keep your old provider for anything we do not cover.