Frequently asked questions
How is the character count calculated?
We charge $10 per 1 million characters, which equals $0.01 per 1,000 characters. The count includes every character in your input text, including spaces and punctuation marks. This transparent, character-based model ensures you only pay for the exact volume of text processed by the API.
Do spaces and punctuation count as characters?
Yes, spaces and punctuation count as characters. Every single character in your input string contributes to the total usage count. This includes standard punctuation marks, spaces between words, and any other symbols included in your text payload sent to the synthesis endpoint.
Is there a minimum amount to top up?
The minimum top-up amount is $5. You can add funds using USDT on the TRON network (TRC-20). There are no monthly fees or subscriptions, so you only pay for what you use, starting with whatever balance you choose to deposit.
How long does it take for USDT to appear?
Credits appear automatically within minutes after your transfer is confirmed. We only accept USDT on the TRON network (TRC-20). Once the blockchain transaction is processed, your account balance updates instantly, allowing you to start generating audio datasets immediately.
Do credits expire if I don't use them?
Your paid balance does not expire. You can top up whenever needed, and the funds remain available indefinitely. The only exception is the free trial credit, which is valid for 7 days after you sign in with Google.
Are failed requests refunded?
Yes, if synthesis fails, the charge is returned in full. You only pay for successful audio generation. If an error occurs during processing, the character count for that request is not deducted from your balance, ensuring you pay only for usable audio.
Can I mix free trial credit with paid balance?
Yes, you can use both. The free trial credit ($0.50, valid for 7 days) is applied first when you sign in with Google. Once the trial expires or is exhausted, your paid balance from USDT top-ups is used automatically.
Is there a monthly fee for API access?
No, there are no monthly fees, subscriptions, or minimum spends. You pay strictly per character used. There are also no seats or enterprise contracts required. You can start with the free trial or top up as needed without recurring costs.
How do I check my current balance?
Visit the /get-api-key/ page on our site. Your dashboard displays your current balance, trial status, characters left, and usage history. You can also see your active API key and a ready-to-use cURL example directly on that same page.
What is the rate limit per API key?
Each API key is limited to 60 requests per minute. If you have not yet topped up, the limit is 10 requests per minute. These limits help manage load; requests are processed sequentially, so high volume may result in a queue.
Does the price change with voice selection?
No, the price is consistent across all voices. Both the 'tts-1' and 'tts-1-hd' models use the same engine and price. You can choose any of the 8 preset voices without any difference in cost per character generated.
Is there a discount for high volume usage?
No, there are no volume discounts or tiered pricing. The rate remains $10 per 1 million characters regardless of usage volume. We offer a flat, transparent rate to simplify billing for product teams and ML engineers generating large datasets.
Can I issue a new API key to reset billing?
No, issuing a new key does not reset your billing or balance. You can issue a new key at any time, which immediately invalidates the old one. Your usage history and character count remain tied to your account, not the key.
Do I pay for text that exceeds 4096 characters?
You pay only for the text actually processed. Each request accepts up to 4096 characters. For longer texts, you split them into parts, and you are charged for each successful part. Failed requests are free, so you only pay for valid audio output.
Cost calculator
Cost per character
Synthetic speech generation is priced purely by input length at a flat rate of $0.01 per 1,000 characters. This means producing one million characters costs exactly $10, regardless of the voice model or audio format selected. The simplicity of character-based billing makes it easy to forecast costs for large-scale dataset creation.
Standard rate
The standard rate applies uniformly across all voice models, including both tts-1 and tts-1-hd. There are no tiered prices for higher fidelity or different voice presets. Every character in your input text contributes equally to the final cost, providing predictable unit economics for every audio file generated.
Character counting
All input characters are counted, including spaces, punctuation marks, and special characters. For example, a 4,096-character request will always bill for exactly 4,096 characters. This consistent counting method ensures that your budget calculations remain accurate whether you are processing short phrases or longer narrative blocks.
No hidden fees
You pay only for the characters sent to the API. There are no charges for API calls, storage, or format conversions. Failed requests are free, ensuring you are not billed for errors. The transparent structure eliminates surprise costs, allowing you to allocate your entire budget to actual data generation.
Trial and credits
Start testing your TTS pipeline immediately without entering payment details. A small credit balance is available to verify voice quality and API integration before committing to larger dataset generation runs.
Understanding your initial trial limits helps you plan how many synthetic samples you can generate during the evaluation phase of your ML project.
Google sign-up credit
Sign up with Google to receive $0.50 of credit. This provides approximately 50,000 characters of generation capacity, enough to test multiple voices and formats.
- Automatic credit application
- No credit card required
Email account trial
Accounts created with email and password do not receive the trial credit. You must top up to begin generating audio, ensuring a clean slate for billing from the first request.
Credit validity
The trial credit is valid for 7 days from account creation.
Payment methods
We accept USDT on the TRON network (TRC-20) for all top-ups. This single payment method keeps the process simple, avoiding credit card fees or complex invoicing workflows that product teams often prefer for bulk data procurement.
USDT TRC-20
Send your top-up via the TRON network using any standard wallet. We do not support other cryptocurrencies or traditional payment rails like PayPal or wire transfers, keeping our infrastructure lean and focused on API reliability.
Minimum top-up
The minimum top-up amount is $5. This low threshold allows teams to test the API’s character-based pricing model with minimal upfront commitment before scaling up for large-scale dataset generation.
Auto-credit
Credits are added to your account automatically within minutes of confirming the blockchain transaction. There is no manual approval process or support ticket required, allowing your ML pipelines to resume generation almost immediately.
Cost examples for datasets
Generating synthetic speech datasets requires precise cost estimation. Our character-based billing ensures you pay exactly for the audio produced, with spaces and punctuation counting toward the total. This transparency simplifies budgeting for product teams and ML engineers sourcing clean audio.
Understanding volume tiers helps in planning large-scale data collection. Below are concrete examples of costs for typical dataset sizes, based on our standard rate of $0.01 per 1,000 characters.
Small dataset costs
A small training set of 100,000 characters costs $1.00. This volume is sufficient for initial model fine-tuning or testing voice consistency across different languages and speeds.
Even with queue processing during high demand, the per-character rate remains constant, ensuring predictable expenses for small batches.
- 100k chars: $1.00
- 500k chars: $5.00
Large scale generation
For larger datasets, such as 10 million characters, the cost is $100.00. This scale is typical for training robust voice models that require diverse phonetic coverage across multiple voices and speeds.
Since failed requests are free, you only pay for successfully generated audio files, protecting your budget from API errors.
- 10M chars: $100.00
- 100M chars: $1,000.00
Budget planning
Start with the $0.50 Google sign-up trial (50,000 characters) to validate your data pipeline before committing to USDT top-ups. This allows you to test voice quality and character counting logic without financial risk.
Track your usage via the dashboard to monitor character consumption in real-time, ensuring you stay within your planned budget for dataset generation.
- Trial: $0.50 (50k chars)

Billing and usage
Billing is strictly metered based on the character count of your input text, including spaces and punctuation. You are charged only for characters that are successfully processed, ensuring you pay exactly for the synthetic audio you receive.
Metered billing
Every request is charged at the standard rate of $0.01 per 1,000 characters. The system calculates the exact character count before synthesis begins, so your cost is predictable even for large-scale dataset generation jobs.
Usage dashboard
Your account dashboard provides a real-time view of your current balance, remaining trial credit, and total characters consumed. You can monitor usage patterns to optimize your dataset generation workflows and track spending without delay.
- Real-time balance updates
- Character count tracking
- Trial credit visibility
Failed request refunds
If a synthesis request fails, the charge is returned in full. You are not billed for characters in audio files that were not successfully generated, ensuring fair billing for every API call.
No subscriptions
VoxyData operates on a pure pay-per-character model, so you only pay for the audio you actually generate. There are no recurring monthly subscriptions tied to your account, meaning your costs scale directly with your dataset volume rather than a fixed calendar fee. This eliminates the risk of paying for unused capacity during quieter weeks.
Our infrastructure supports flexible usage patterns without administrative overhead. Whether you are generating thousands of hours of training data or running a small pilot, your billing adjusts instantly to your actual consumption. This transparency helps ML teams forecast exact project costs without hidden tier jumps or surprise renewal charges.
No monthly plans
You are not locked into a recurring billing cycle or a fixed quota that expires if unused. If your dataset generation pauses, your bill drops to zero immediately. This structure is ideal for teams with variable workloads, such as those building voice models for specific products or testing new synthetic voice pipelines.
There are no annual commitments or auto-renewing contracts. You can top up your USDT balance once and let it sit until needed. This simplicity reduces accounting overhead for finance teams and ensures that every dollar spent goes directly toward character synthesis, not platform access fees.
No seat limits
Your API key is the only requirement for access, not the number of engineers or developers using it. Multiple team members can share the same key to generate datasets simultaneously without incurring extra per-seat fees. This is particularly useful for ML teams where several data scientists might be running parallel generation jobs from different scripts.
Since VoxyData is a simple API rather than a collaborative workspace, there is no need to manage user roles or invite colleagues to a dashboard. Your usage and billing are tied to the key and account balance, allowing your entire engineering team to focus on model training rather than license management.
- Share one API key across your entire team
No minimum spend
You can start with the free trial credit or a small top-up of just $5. There is no requirement to commit to a large monthly volume before you see value. This low barrier to entry allows product teams to validate synthetic audio quality for their ML models without significant upfront financial risk.
As your dataset grows, you can increase your USDT deposits incrementally. There are no penalties for low-volume months or bonuses for high-volume months. The pricing remains consistent at $0.01 per 1,000 characters regardless of whether you generate 100 characters or 100 million, ensuring predictable unit economics for your data projects.
Rate limits and throughput
Understanding throughput is essential when generating large synthetic speech datasets. While the API supports rapid character generation, request volume is capped to ensure consistent voice quality and stable processing.
Requests are processed sequentially, meaning concurrent calls may experience slight delays during peak usage. This behavior is normal and ensures reliable audio synthesis without dropping data.
Requests per minute
Each API key is limited to 60 requests per minute after the initial top-up. Before your first deposit, the limit is lower at 10 requests per minute.
This limit applies per key, not per account. If you generate multiple keys, the total throughput scales linearly with the number of active keys.
- Standard limit: 60 requests/minute/key
- Pre-top-up limit: 10 requests/minute
- Limits reset every 60 seconds
Queue processing
Requests are processed one after another. Under load, a request may wait in a queue before synthesis begins.
Generation speed is roughly real-time, averaging about 25 characters of text per second of audio. Short phrases typically take a few seconds to complete.
- Sequential processing ensures stability
- Average speed: 25 chars/sec
- Short phrases may take a few seconds
Scale considerations
For large-scale dataset generation, consider splitting texts into parts of up to 4096 characters. This allows you to manage request volume effectively.
If you need higher throughput, issue additional API keys. Each key operates independently, allowing you to parallelize requests across multiple threads or services.
- Max input: 4096 characters per request
- Split long texts for better control
- Use multiple keys for parallel generation
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This helps prevent overfitting and improves model generalization.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- openai-python on GitHub
- LibriTTS-R corpus (OpenSLR 141)
- W3C WAI: making audio and video accessible
- LibriTTS-R paper (arXiv 2305.18802)
- ISO 639 language codes
- Tether (USDT)
- LibriVox public-domain audiobooks
- MDN: Web Speech API
- openai-node on GitHub
- OpenAI API reference: create speech
- Creative Commons Attribution 4.0
