Text-to-Speech Cost Calculator
Inputs
| Character Count | 10,000 |
|---|---|
| Price (per 1M characters) | 15 $ |
Text-to-Speech Cost Calculator
Estimate the cost of synthesizing speech with a text-to-speech API from the character count and the provider’s per-million-character rate, in any currency.
Inputs
Synthesis
Results
Enter a value to see results.
Text-to-speech cost explained
The text-to-speech cost of a speech-synthesis API is the price charged to turn written text into spoken audio, computed from the number of input characters and the provider's published per-million-character rate. Anyone producing voiceovers, audiobook drafts, accessibility narration, or spoken assistant replies needs this figure, because synthesis is metered by characters of input text rather than by the length of the audio it produces.
How synthesis is billed
Text-to-speech providers count the characters in the text you submit — letters, spaces, and punctuation all included — and bill per one million characters. This is the mirror image of transcription, which bills by audio duration. It also differs from text language models, which bill by tokens. Because a single character costs a tiny fraction of a cent, providers publish the rate per million characters to keep the headline number readable.
The formula
With input characters at a price per million characters, the cost is
C=106n⋅pThe division by converts the per-million quote into a per-character rate before multiplying by the character count. The relationship is linear: doubling the text doubles the cost.
Worked example
Consider a narration script of 8,400 characters synthesized at $15 per million characters, a published rate for a higher-quality voice tier. The cost is
C=1068400×15=0.0084×15=$0.126so the script costs about thirteen cents to voice. A 50,000-word audiobook draft — roughly 300,000 characters — would cost around $4.50 at the same rate, which makes synthesis practical for iterating on long-form narration.
Notes and variations
The rate depends on the voice tier. A standard tier costs less per character than a high-fidelity tier that produces more natural prosody, so set the rate to the tier you intend to use. Some providers count a multi-byte character as one character and others count its bytes, which matters for non-Latin scripts; check the counting rule before budgeting in those languages. Re-synthesizing the same text after an edit is billed again at the full character count.
Application
Per-script cost is the building block for narration budgets — multiply it by the number of scripts to estimate recurring spend. For the reverse direction — transcribing audio into text rather than synthesizing it — see the Audio Transcription Cost Calculator. When synthesis is paired with a language model that drafts the text, the prompt and completion side is handled by the Token Cost Calculator.
Frequently Asked Questions (FAQ)
Is speech billed by tokens or characters?
Text-to-speech is billed by input characters, not tokens. Every character of the text you send — letters, spaces, and punctuation — counts toward the total. This differs from text models, which bill by tokens. As a rough guide, English text averages a little under six characters per word including the trailing space.
Why do higher-quality voices cost more?
Providers usually offer a standard tier and a higher-fidelity tier that produces more natural prosody and clearer audio. The high-fidelity tier is billed at a higher per-character rate because it uses a larger model. Switch the per-million rate to the tier you intend to use; the character count stays the same.
How can I estimate the character count?
The character count is simply the length of the text string, so most editors and word processors report it directly. If you only know the word count, multiplying by roughly six gives a workable estimate for English, since the average word plus its following space is close to six characters.
Disclaimer
This is a cost estimate. Published text-to-speech rates change over time and vary by provider and voice tier. Verify current pricing and the exact character-counting rules with your provider before relying on a figure for budgeting.