Frequently asked questions
Is this a voice cloning service?
No. We do not offer custom voice cloning for customers, nor do you upload your own voice. The service uses eight fixed presets based on licensed recordings. If you need a specific voice, you must select from the available options like Aria or Atlas. Custom cloning is not part of our feature set.
Do you clone my voice for audiobooks?
No. We do not clone your voice. Our audio is generated from eight preset voices derived from LibriTTS-R and LibriVox recordings. You cannot upload a sample to create a custom clone. Each generated file carries an inaudible watermark marking it as synthetic speech, distinct from a cloned human voice.
Which voices sound best for fiction?
Choose based on your narrative style. Luna offers a soft, storytelling tone ideal for gentle fiction. Aria provides a warm, clear narrator voice. For energetic or promotional fiction, Felix works well. Sage is suitable for instructional or calm narratives. You can test each voice via the /v1/voices endpoint to find the best fit.
How do I handle long manuscripts?
Split your text into parts of up to 4096 characters per request. Send each part via POST to the speech endpoint, then join the resulting audio files. This ensures reliable processing. Longer texts take roughly real-time to generate, so plan your workflow accordingly. Use the language parameter if needed for non-English scripts.
What happens if the API fails?
Failed requests are free. If synthesis fails, the charge is returned in full. You can retry the request without additional cost. This applies to any failure during the audio generation process. Your balance remains unaffected by unsuccessful attempts, ensuring you only pay for successfully generated audio files.
Are the voices based on real people?
Yes. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. While the voices originate from real speakers, they are processed into presets. Each generated file includes an inaudible watermark to identify it as synthetic speech, distinguishing it from the original source recordings.
Do I need to credit the voice actors?
The facts do not specify a mandatory credit requirement for end-users. The voices are based on LibriTTS-R (CC BY 4.0) and LibriVox. Check the specific license terms for LibriVox if required by your distribution channel. Generally, the inaudible watermark marks the audio as synthetic, but explicit attribution depends on your platform's rules.
Can I use the audio commercially?
Yes. The product is a cloud API for generating audio files. The voices are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. You receive the audio file directly. There are no restrictions mentioned on commercial use of the generated output, provided you comply with the underlying source licenses.
What is the character limit per request?
The input text limit is 4096 characters per request. This includes spaces and punctuation. For longer manuscripts, you must split the text into chunks of this size. Each chunk generates a separate audio file that you can join later. This limit applies to every individual API call you make.
Do you support SSML for prosody?
No. We do not support SSML. You cannot use prosody tags or custom pronunciation lexicons. The API accepts plain text input. If you need specific pacing, adjust the speed parameter between 0.25 and 4.0. The engine processes text linearly without additional markup instructions for prosody control.
Is there a free trial for freelancers?
Yes. Signing in with Google provides $0.50 of credit, valid for 7 days. This covers about 50,000 characters. Accounts created with email and password start without trial credit. You can top up via USDT on the TRON network later. The trial allows you to test the API before purchasing.
How long does generation take?
Generation takes roughly real time, about 25 characters per second for long texts. Short phrases may take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue. Plan your workflow to account for potential delays during high-traffic periods.
Can I mix voices in one project?
Yes. Generate audio for each section using the desired voice preset. Then join the resulting audio files. The API does not support mixing voices in a single request. You manage the sequence by sending multiple POST requests with different voice parameters and combining the outputs yourself.
Do you offer refunds for errors?
Yes. Failed requests are free, and the charge is returned in full. If synthesis fails, you are not billed. This ensures you only pay for successfully generated audio. Refunds are automatic for failed attempts, so your balance is restored immediately.
First request
curl
curl https://api.60w.info/v1/audio/speech \
-H "Authorization: Bearer $TTS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
--output hello.mp3Why Freelancers Choose API-Based Audiobooks
API-based generation removes the guesswork from audiobook production. You control the exact pacing and speed of every sentence, ensuring the final audio matches your creative vision without waiting for external scheduling.
Control over pacing and speed
Adjust speed from 0.25x to 4.0x to fit your narrative style. This precision helps maintain consistent tone across chapters, which is critical for self-published authors managing their own production pipelines.
No monthly subscriptions
Pay only for what you use. With transparent character-based pricing, there are no hidden volume fees or unused seats. Your balance does not expire, allowing you to pause production between books without losing credits.
Instant generation workflow
Send text via API and receive audio files immediately. This direct integration allows you to generate chapters on demand, streamlining the path from manuscript to finished audiobook without intermediate human delays.
The Voice Cloning Context
Voice cloning typically involves training a model on hours of specific voice recordings to replicate a unique timbre. Our API uses eight distinct presets based on licensed recordings, offering consistent character voices without the need for custom training data or uploads.
Each voice is a fixed artistic preset, not a clone of a specific celebrity or individual. This distinction ensures predictable audio quality and clear licensing terms for your self-published audiobooks.
What voice cloning actually is
Cloning requires feeding a model large datasets of a target voice to learn its unique speech patterns and tone. This process is resource-intensive and often requires explicit consent from the voice owner.
You select from eight pre-defined voices that are ready to use immediately via the API.
How our preset voices differ
The eight voices (Aria, Nova, Luna, Sage, Felix, Milo, Atlas, Orion) provide distinct tonal ranges suitable for various storytelling styles, from warm narration to energetic promos. They are not clones but distinct audio characters.
These presets are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain LibriVox readings, ensuring a reliable foundation for commercial use.
- Based on LibriTTS-R and LibriVox sources
- Eight distinct character presets
- No custom training required
Consent and licensing explained
The underlying recordings are licensed under CC BY 4.0 or are public domain, providing a clear legal basis for using these voices in your audiobooks. You do not need to negotiate with individual voice actors.
Generated audio files contain an inaudible watermark indicating they are synthetic, which helps distinguish them from human-recorded tracks in distribution channels.
8 Licensed Voices for Storytelling
AudioPilot provides eight distinct voice presets designed for clear, engaging audiobook narration. Each voice offers a specific tonal quality, allowing you to match the narrative style of your manuscript precisely.
Female presets: Aria, Nova, Luna, Sage
Aria delivers a warm, clear narration suitable for general fiction. Nova offers a bright, upbeat tone for light reads. Luna provides a soft, storytelling voice ideal for emotional scenes. Sage is calm and instructional, perfect for non-fiction or self-help texts.
Male presets: Felix, Milo, Atlas, Orion
Felix brings energetic, promo-style delivery for fast-paced narratives. Milo offers a friendly, conversational tone. Atlas provides a neutral, news-like delivery for straightforward content. Orion features a deep, documentary voice for serious or historical texts.
Source recordings from LibriVox
All voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. This ensures a clear licensing path for your generated audio files.
- Base: LibriTTS-R (CC BY 4.0)
- Source: LibriVox public-domain readings
- Clear licensing for commercial use
Supported Languages and Scripts
AudioPilot supports 23 languages, covering major global markets including English, Spanish, French, German, Japanese, and Arabic. Each voice preset is capable of speaking any of these languages, giving narrators flexibility without needing to hire separate voice talent for different regions.
23 languages supported
We support Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. Notable exclusions include Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, and Bengali.
Automatic language detection
The API guesses the language from the text script. Latin-script text defaults to English if unspecified. This works well for single-language manuscripts but may misidentify mixed-language passages.
Setting language explicitly
For non-English Latin-script text, always set the language parameter to the correct ISO code. This ensures accurate pronunciation and intonation, preventing the system from defaulting to English for words like "café" or "naïve".
Technical Requirements for Authors
Authors must break manuscripts into segments of 4,096 characters or fewer for each API request. This limit ensures reliable synthesis and prevents truncation errors during generation.
For longer works, split the text sequentially, generate each part, and join the audio files. Short phrases may take a few seconds to process, so account for queue time under load.
4096 character input limit
Each POST request accepts a maximum of 4,096 characters for the input field. Spaces and punctuation count toward this total. Exceeding the limit will cause the request to fail.
- Count characters carefully before sending.
- Include spaces and punctuation in your count.
- Split texts slightly under 4,096 to avoid edge-case errors.
Splitting long manuscripts
Divide your manuscript into logical chunks, such as chapters or scenes. Send each chunk as a separate request to the API. Download the resulting audio files and concatenate them using standard audio editing software or a simple script to create the final audiobook.
Note that requests are processed sequentially. Under load, a request may wait in a queue before synthesis begins.
- Maintain consistent pacing across segments.
- Use a consistent voice for all parts.
- Join files using standard tools.
Audio formats: MP3, WAV, FLAC
Select your preferred response_format when making the API call. MP3 offers small file sizes for easy distribution. WAV provides lossless quality for editing. FLAC offers lossless compression for archival storage. PCM is also available for raw data processing. All formats support the same 8 voices and 23 languages.
The API returns the audio file directly in the response body. No additional storage is required unless you choose to save it yourself.
- MP3: Suitable for web distribution.
- WAV: Suitable for professional editing.
- FLAC: Lossless compression for archives.

Transparent Pricing Model
The pricing model is straightforward: you pay $10 for every 1 million characters processed. This rate applies uniformly to both tts-1 and tts-1-hd models, ensuring consistent costs regardless of the quality tier you select for your audiobook chapters.
$10 per million characters
Each character, including spaces and punctuation, counts toward your usage. At this rate, 1,000 characters cost just $0.01. This allows you to estimate production costs precisely based on your manuscript length before starting synthesis.
Failed requests are free
If a synthesis request fails due to server issues or invalid parameters, you are not charged. Any attempted charge is returned in full, so you only pay for successfully generated audio files.
No hidden volume fees
There are no monthly subscriptions, per-seat fees, or minimum spend requirements. You top up your balance with USDT (TRC-20) and spend only what you use. Your balance does not expire, allowing flexible production schedules.
Data Privacy and Training
Your input text is processed exclusively to generate the audio file and is not retained for model training. This ensures your unpublished manuscripts remain private during the synthesis process. Once the audio is returned in the response, the source text is not stored.
We maintain request metadata—such as timestamp, character count, and status—strictly for billing and usage tracking. This data helps us accurately calculate your charges based on the $10 per million characters rate without retaining the actual content of your stories.
Text used only for generation
The API processes your text solely to produce the audio output. We do not use your input data to improve the voice models or train future versions of the TTS engine. Your content remains yours.
No training on input data
Unlike some consumer-grade TTS services, we do not ingest your submitted text into our training datasets. This is particularly important for self-publishing authors who want to ensure their unpublished works are not used to train the underlying AI models.
Metadata kept for billing
We store request metadata including time, character count, and status. This information is used solely to calculate your usage charges and verify that failed requests are refunded correctly. The actual text content is not stored after generation.
Getting Started in Minutes
Create your account on the /get-api-key/ page using either Google sign-in or a standard email and password. This immediate registration gives you access to your personal dashboard, where you can manage credentials and monitor usage metrics in real time.
Once registered, your unique API key is generated instantly. You can use this key to authenticate requests to the audio endpoint. If you need a fresh key, you can issue a new one at any time, and the previous one stops immediately.
Sign up with Google or email
Registering takes seconds. Google sign-in grants a $0.50 trial credit valid for 7 days, while email accounts start without trial credit. No credit card is required to begin.
- Google or email registration
- Instant access to dashboard
Get your API key instantly
Your key appears on the dashboard immediately after sign-up. Copy it to your clipboard or note it securely. You can regenerate it anytime from the same page.
- Key visible on dashboard
- Regenerate anytime
Use our curl example
Start with our ready-to-use curl command. Replace the placeholder with your key and send a POST request to the API base. Receive your audio file directly in the response.
- Pre-written curl command
- Direct API integration
Common Misconceptions About TTS
Many authors confuse preset voices with custom cloning. Our eight voices are based on licensed LibriTTS-R recordings, not cloned from specific celebrities. You select a preset that fits your narrative style, ensuring consistent, clear narration without the complexity or legal ambiguity of custom voice models.
It is not real voice cloning
We do not offer custom voice cloning. You cannot upload a sample to create a unique voice profile. Instead, you choose from eight distinct presets designed for storytelling, ranging from warm narrators to deep documentary styles. Each voice is a fixed asset, ensuring predictable quality across your entire audiobook.
- No celebrity impersonations or custom voice uploads.
Watermarked synthetic audio
Every audio file generated by our API contains an inaudible digital watermark. This mark identifies the file as synthetic speech, which is standard for AI-generated content. It does not affect audio quality or playback but provides a clear distinction between human-recorded and AI-narrated content.
Not suitable for real-time streaming
Our API is designed for batch processing, not real-time streaming. You send a text request and receive the complete audio file once synthesis is finished. While fast, this is not a live streaming solution like WebSocket. For audiobooks, this batch model is ideal as it allows you to generate entire chapters efficiently.
- Batch processing only; no WebSocket streaming.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Once you receive the audio files for each chunk, concatenate them in the correct order.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- ISO 639 language codes
- Node.js documentation
- Creative Commons Attribution 4.0
- FFmpeg documentation
- openai-python on GitHub
- W3C WAI: making audio and video accessible
- curl documentation
- TRON developer hub: TRC-20
- OpenAI API reference: create speech
- LibriTTS-R paper (arXiv 2305.18802)
- LibriVox public-domain audiobooks
