Frequently asked questions
What is voice cloning exactly?
Voice cloning creates a synthetic voice model from a sample recording. Our API does not offer custom voice cloning for customers, nor does it allow uploading your own voice or designing new voices. You use one of the eight preset voices instead.
Is voice cloning the same as TTS?
No. Text-to-speech converts text into audio using preset voices like Aria or Felix. Voice cloning generates a unique voice model from a specific person’s recording. We provide TTS with eight presets; we do not offer the cloning technology to create custom voice models.
Do I need to train a model for cloning?
Since we do not offer custom voice cloning, you do not need to train a model. You simply select one of the eight available preset voices in the API request. Each preset is based on licensed recordings and ready for immediate use.
How much does voice cloning cost?
We do not offer voice cloning, so there is no specific cloning cost. Standard TTS costs $10 per 1 million characters. Failed requests are free. You pay only for the audio generation using our preset voices, with no minimum spend or monthly plans.
Can I clone a celebrity's voice?
We do not offer custom voice cloning, so you cannot clone a celebrity’s voice. You can only use the eight preset voices provided. These are based on licensed recordings from LibriTTS-R and LibriVox, not on specific celebrities or public figures.
Do I need consent to clone a voice?
Custom voice cloning is not offered, so consent rules for cloning do not apply. Our preset voices are based on licensed recordings. Generated files carry an inaudible watermark marking them as synthetic speech. You use the provided voices under our standard terms without additional licensing steps.
How is cloning different from presets?
Cloning creates a unique voice model from a specific person’s recording. Presets are fixed voices like Aria or Felix, based on licensed audiobook readings. We offer presets only. You cannot upload audio to create a clone, so you must choose from the eight available options.
Are cloned voices more realistic?
We do not offer cloned voices, so this comparison is not applicable. Our preset voices are based on licensed recordings and include an inaudible watermark to mark them as synthetic. They provide clear, natural-sounding speech for blogs and WordPress posts without the complexity of custom models.
Can I use cloned voices commercially?
We do not offer cloned voices, so commercial cloning usage does not apply. You can use our preset voices commercially. The API charges $10 per 1 million characters. Generated audio is returned in the response and not stored, allowing flexible use for your content.
Does the API support custom cloning?
No. The API supports eight preset voices: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion. You cannot upload a voice sample to create a custom clone. You send text via POST request and receive audio generated by one of these fixed presets.
What is a voice model?
A voice model is a synthetic representation of a specific person’s voice, typically created through cloning. We do not offer custom cloning or voice models. We provide eight preset voices based on licensed recordings. You select a preset voice ID in your API request to generate audio.
How long does cloning take?
Cloning is not offered, so there is no cloning training time. Standard TTS generation takes roughly real time, about 25 characters per second. Short phrases take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue.
Defining Voice Cloning
Voice cloning creates a unique digital replica of a specific person’s voice. Unlike standard text-to-speech systems that use preset models, cloning requires capturing distinct vocal characteristics to build a custom audio profile.
Creation of a unique voice model
This process generates a distinct voice model tailored to individual speech patterns. It allows for highly personalized audio output that mimics a specific speaker rather than using a generic preset.
Requires training data
Accurate cloning depends on substantial training data. The system analyzes existing recordings to learn nuances, requiring significant input before the model is ready for synthesis.
Higher cost and complexity
Cloning involves higher costs and technical complexity compared to standard TTS. Users must manage data collection and processing, whereas preset voices offer immediate, straightforward generation without extra setup.
Standard TTS vs Cloning
Standard text-to-speech relies on eight preset voices like Aria or Felix, which are ready to use immediately. Voice cloning builds a unique model from your specific audio samples, offering a custom sound but requiring a training phase. While cloning creates a one-of-a-kind voice, presets ensure consistent, predictable quality across all your content.
Cloning typically involves higher costs and technical complexity due to the training data required. Standard TTS generates audio instantly using established models, making it ideal for bloggers who need rapid content production without managing custom voice assets or waiting for model training to complete.
Preset voices vs custom voices
Presets provide eight distinct voices (Aria, Nova, Luna, Sage, Felix, Milo, Atlas, Orion) that speak any supported language, including Spanish. You select a voice name in your API request. Custom cloning requires uploading specific audio samples to train a new voice model, which is not currently offered by this service.
Using presets ensures your audio sounds consistent across all blog posts. Cloning aims for a unique brand voice but may vary in quality depending on the source audio. Presets are licensed from LibriVox recordings, ensuring clear rights for commercial use.
- Eight preset voices available instantly
- No training data required for presets
- Custom voices need specific audio samples
Instant generation vs training time
Standard TTS generates audio files in seconds. Long texts process in roughly real-time speed. Cloning requires a training period before the new voice can be used. This delay adds time to your workflow, whereas presets are immediately available after API key setup.
For WordPress bloggers, instant generation means you can create voiceovers while editing posts. There is no waiting for model training. You simply send text to the API and receive an audio file, keeping your content pipeline smooth and uninterrupted.
Cost differences explained
Standard TTS costs $0.01 per 1,000 characters, with no minimum spend or monthly fees. Failed requests are free. Cloning typically incurs higher upfront training costs and potentially higher per-minute rates for usage, depending on the provider. This service offers predictable pricing for presets.
For blogs generating frequent posts, the flat character rate makes budgeting easy. You pay only for what you use. There are no seats or hidden fees. This simplicity suits no-code creators who need reliable audio without complex financial commitments.
Spanish Language Support
VocalPress supports Spanish with eight preset voices, ensuring your blog content sounds natural and engaging without requiring complex setup. Each voice handles Spanish fluently, from the warm narration of Aria to the energetic promo style of Felix.
These voices deliver natural pronunciation across all 23 supported languages, including Spanish, so you can create high-quality audio posts instantly. There is no need for specific Spanish cloning to achieve clear, professional results for your WordPress site.
8 voices speaking Spanish
All eight available voices can speak Spanish effectively. You can choose from female voices like Aria and Nova or male options like Felix and Milo, depending on the tone of your blog post.
Each voice brings a distinct character, whether you need a calm instructional tone with Sage or a deep documentary style with Orion. This variety helps you match the audio to your content's mood.
- Aria, Nova, Luna, Sage
- Felix, Milo, Atlas, Orion
Natural pronunciation in presets
The presets handle Spanish phonetics accurately, avoiding the robotic feel common in basic TTS. Text is converted to speech with clear articulation, making it suitable for reading articles or narrating stories.
You can adjust the speed between 0.25 and 4.0 to fit your audience's listening preferences. This flexibility ensures the audio matches the pacing of your written content perfectly.
No need for specific Spanish cloning
Since the presets already handle Spanish natively, you do not need to invest in custom voice cloning. This saves time and money, allowing you to focus on creating content rather than managing voice models.
Simply select the voice that fits your brand and generate the audio via the API. The result is consistent, high-quality speech ready for embedding in your WordPress posts.
Consent and Licensing
Every vocal performance in our engine originates from licensed recordings sourced directly from LibriVox. This foundation ensures that the voices you hear are not random synthetic constructs but are built upon authentic human speech, providing a natural and engaging listening experience for your WordPress readers.
The underlying audio data is covered by a CC BY 4.0 license. This means the foundational recordings are freely available for reuse, allowing us to offer a reliable, cost-effective TTS solution.
Voices from LibriVox recordings
Our eight preset voices—Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion—are derived from public domain audiobook readings. These recordings serve as the training basis for our synthesis engine, ensuring that every generated segment retains the organic rhythm and tone of human narration.
By leveraging these established recordings, we maintain high audio fidelity. The voices sound natural because they are rooted in real speech patterns, not just algorithmic approximations. This approach delivers consistent quality across all 23 supported languages.
- Sources: LibriVox public domain audiobooks
- Basis for all 8 preset voices
- Ensures natural speech rhythm
CC BY 4.0 license
The CC BY 4.0 license governs the usage of the source recordings. This Creative Commons license allows for commercial use and modification, provided appropriate credit is given. This legal framework supports our business model and enables us to distribute high-quality audio to our users.
Because the source material is clearly licensed, there are no hidden royalties or complex clearance processes for the voice models themselves. You can generate audio for your blog posts with confidence, knowing the underlying voice assets are legally secure for commercial distribution.
Public domain source material
The original recordings come from volunteers reading public domain books, meaning the copyright on the underlying text and the specific recordings has expired or been waived. This public domain status is crucial for maintaining low costs and avoiding perpetual licensing fees.
This transparency extends to the generated audio. While the voices are synthetic, their lineage is clear. We do not claim ownership over the original human voices, only the specific synthesis and processing we apply to create your final audio files.
Watermarking Synthetic Audio
Every audio file generated by VocalPress contains an inaudible watermark that explicitly marks the content as synthetic speech. This digital signature is embedded directly into the audio data, ensuring that listeners can be aware of the artificial nature of the voice without it affecting the listening experience.
This feature provides clear transparency for your audience, distinguishing your AI-generated voiceovers from human-recorded narration. It helps maintain trust with your WordPress readers by being upfront about the technology used, without requiring any additional metadata or visible labels on your blog posts.
Inaudible watermark in files
The watermark is embedded at a level that does not interfere with audio quality or playback. It remains consistent across all formats, including MP3, WAV, OPUS, FLAC, and PCM, ensuring reliable detection regardless of how the audio is distributed or played back on your site.
- Embedded in all output formats
- Does not degrade audio quality
- Persists through standard playback
Marks audio as synthetic
Unlike voice cloning which mimics a specific person, our standard TTS voices are distinct synthetic entities. The watermark confirms the voice is AI-generated, not a recording of a real human, which is crucial for transparency in content creation and media reporting.
- Indicates AI origin
- Distinguishes from human voice
- No visual cue needed
Transparency for listeners
By marking audio as synthetic, you offer honesty to your audience. This is particularly useful for blogs where clarity about content sources is valued. It aligns with growing expectations for labeling AI-generated media without disrupting the user experience.
- Builds audience trust
- Clear content labeling
- Supports media transparency
When to Use Cloning vs Presets
Presets like Aria and Felix provide consistent, high-quality Spanish audio instantly. This reliability is ideal for bloggers who need to generate voiceovers for multiple posts without waiting for training. The uniform quality ensures your audience hears the same clear tone across all your content.
Voice cloning creates a unique model from your own recordings. It is best for brands needing a distinct, proprietary voice that presets cannot replicate. However, it requires more data and processing time, making it less suitable for rapid, high-volume blog updates.
Presets for consistency and speed
Presets deliver immediate results with uniform quality. This is crucial for maintaining a steady publishing schedule. You can generate audio for a 4,000-character post in seconds, ensuring your workflow remains uninterrupted by long training periods.
- Instant generation for rapid content scaling
- Uniform tone across all blog posts
- No data collection or training required
Cloning for unique brand voices
Cloning captures the specific nuances of your own voice. It is valuable if your brand identity is deeply tied to your personal voice. This option is less about speed and more about creating a distinct auditory signature that cannot be duplicated by standard presets.
- Captures unique vocal nuances
- Ideal for personal brand blogs
- Requires significant training data
Cost-benefit analysis for blogs
Presets are cost-effective for most blog needs. At $0.01 per 1,000 characters, you pay only for usage. Cloning involves additional setup costs and complexity. For most WordPress bloggers, the speed and low cost of presets offer a better return on investment than maintaining a custom cloned model.
- Low per-character cost
- No upfront cloning fees
- Simpler integration for standard workflows
Legal Considerations for Blogs
Understanding the legal landscape ensures your blog remains compliant while using synthetic audio. Since our voices are derived from public domain LibriVox recordings under a CC BY 4.0 license, you inherit clear usage rights for the generated speech.
The generated audio files are yours to use immediately upon creation. There are no complex licensing agreements or royalty payments required beyond the API usage cost, making it straightforward for content creators to focus on their narrative.
Usage rights of generated audio
You retain full rights to the audio files generated through the API. The input text is used solely for synthesis and not for training, ensuring your specific content remains private. The audio files can be embedded in posts, used in podcasts, or distributed via RSS feeds without additional legal hurdles.
Since the base voices originate from public domain sources, there are no restrictions on how you distribute the resulting audio. You can use the audio in both free and paid content without worrying about additional licensing fees or royalty disputes.
- Audio files are generated on-demand and returned instantly.
- Input text is not used for training models.
- No ongoing royalty payments are required.
Attribution requirements
The CC BY 4.0 license for the source recordings generally requires attribution. However, since you are using the audio as a final product rather than redistributing the raw voice model, standard blog attribution practices apply. You may credit VocalPress or note the use of AI audio if desired, but it is not strictly mandated for every single post.
For most blog contexts, simply stating that audio is AI-generated is sufficient transparency. Check the specific license of the source recording if you plan to redistribute the raw voice model, but for standard web audio, standard copyright norms apply.
- CC BY 4.0 applies to source recordings.
- Standard blog attribution is recommended.
- No complex attribution chains needed.
Commercial use permissions
The generated audio can be used for commercial purposes, including monetized blogs, sponsored content, and paid newsletters. There are no restrictions on using the voiceovers for profit. The only requirement is that you have paid for the character usage via the API.
You can embed the audio in YouTube videos, podcast apps, or e-learning modules without fear of copyright claims. The synthetic nature of the audio does not limit its commercial applicability, provided the underlying license allows it, which it does for public domain sources.
- Audio can be used in monetized content.
- No restrictions on paid newsletters.
- Suitable for sponsored posts and ads.

Technical Implementation
We do not clone voices for you; you select from eight preset voices that are ready to use immediately. This eliminates the waiting period and technical overhead of training custom models, allowing you to generate audio files instantly via our API.
No code cloning required
Voice cloning involves training a unique model on specific audio data, which adds cost and complexity. We skip this step entirely. You simply choose from our existing library of eight licensed voices, ensuring consistent quality without waiting for synthesis training.
API-driven voice selection
Your workflow relies on a simple POST request to our endpoint. You specify the voice parameter (e.g., aria, nova, or atlas) alongside your text and language code. The API returns the audio file directly, handling all synthesis logic behind the scenes.
- Use standard HTTP requests or the OpenAI SDK.
- Select from 8 presets: aria, nova, luna, sage, felix, milo, atlas, orion.
- Specify language explicitly (e.g.,
es) for accurate Spanish pronunciation.
Easy integration for WordPress
For WordPress bloggers, this API fits into your existing no-code or low-code workflows. You can generate voiceovers programmatically or via simple scripts, avoiding complex plugins. The straightforward JSON response allows you to embed audio files directly into your posts or sidebar widgets.
- Works with standard WordPress PHP or JavaScript integrations.
- No custom voice cloning needed for brand consistency.
- Instant generation supports rapid content publishing cycles.
Audio Quality Comparison
VocalPress presets deliver consistent, high-fidelity audio across all 23 supported languages, including Spanish. Each voice is engineered for clarity, ensuring your blog content remains professional and easy to digest without the variability inherent in custom training processes.
While custom cloning can mimic specific tones, quality often depends on input data. In contrast, VocalPress provides uniform, polished output instantly. This reliability means your WordPress posts maintain a consistent audio brand, regardless of which preset you select.
Presets offer high quality
Our eight preset voices—such as Aria or Felix—are trained on licensed LibriVox recordings, ensuring clear, natural-sounding speech. They handle complex sentence structures and punctuation smoothly, making them ideal for reading long blog posts or short updates alike.
Users benefit from predictable audio characteristics. There is no risk of artifacts or unnatural pauses common in less refined cloning models. The output is immediately usable for embedding directly into your WordPress posts or downloading for offline use.
- Consistent tone across all requests
- No training data quality issues
- Ready for immediate commercial use
Cloning quality varies
Custom voice cloning relies heavily on the quality and quantity of source audio. Variations in background noise or recording conditions in the source material can lead to inconsistent results in the final cloned voice.
Unlike presets, which offer standardized excellence, cloned voices may exhibit quirks or artifacts depending on the training data. For most bloggers, the reliability of a preset voice outweighs the niche appeal of a cloned voice that might sound inconsistent.
Real-time generation advantage
VocalPress generates audio in real-time, with typical speeds matching natural speech patterns. This allows for immediate feedback and editing. You do not wait hours for a cloning model to train; you simply send text and receive audio.
This speed supports rapid content creation workflows. If you need to adjust a blog post’s tone, you can switch voices or regenerate text instantly without the overhead of retraining a model, keeping your publishing schedule on track.
Scalability for Content Creators
VocalPress enables content creators to scale audio production without bottlenecks. By generating many voices instantly, you can produce full blog post voiceovers in seconds rather than minutes. The API handles requests sequentially, so long texts are split and joined automatically, maintaining a steady output rate of roughly real-time for extended content.
Generate many voices instantly
Each request processes up to 4096 characters, allowing you to batch-produce episodes or series efficiently. With a rate limit of 60 requests per minute, you can automate your publishing workflow without manual intervention. This speed suits WordPress bloggers who need to publish multiple posts daily without waiting for human narrators or rendering queues.
No waiting for training
Unlike custom cloning, there is no waiting period for model training. You select from eight preset voices and start generating audio immediately. This instant availability means you can respond to trending topics or update old posts with new audio without scheduling delays or production backlogs.
Consistent brand voice
Maintaining a consistent brand voice is simpler with presets. You can assign specific voices to different content categories, such as using Atlas for news or Luna for storytelling. This consistency builds listener recognition across your site, ensuring that your audio identity remains stable as your content library grows.
Common Misconceptions
Many creators assume voice cloning is simply pitch-shifting or speed-adjusting a recording. It is actually a distinct process that analyzes and replicates vocal characteristics to build a new model. Our standard TTS uses fixed presets like Aria or Felix, which do not require this extra training step.
Cloning is not just voice changing
Voice changing alters an existing audio file, while cloning generates new speech from text using a learned model. Our API delivers instant speech without the need for voice changing or post-processing.
Presets are not robotic
Presets like Nova or Orion are based on licensed human recordings from LibriVox. They sound natural and conversational, avoiding the monotone quality often associated with older TTS systems. You get human-like clarity without waiting for training.
Cloning does not guarantee likeness
Cloning creates a synthetic approximation rather than an exact copy. The result depends on the training data and may not perfectly match the original speaker’s unique nuances. Presets offer consistent, reliable results every time.
Future of TTS and Cloning
Synthetic media is shifting from novelty to utility, with text-to-speech becoming a standard component for content creators who need scalable audio without production overhead. As models improve, the line between preset voices and cloned audio blurs, though our API focuses on reliable, instant generation using licensed recordings rather than complex custom training.
Trends in synthetic media
The industry is moving toward higher efficiency and lower latency. Users now expect immediate audio generation rather than waiting for voice models to train. This shift favors preset voices that offer consistent quality and speed, aligning well with the needs of bloggers and no-code creators who publish frequently.
- Shift from custom training to instant preset generation
- Focus on efficiency and real-time output
- Integration of synthetic audio into standard workflows
Regulatory changes
As synthetic audio becomes more common, regulations regarding transparency and consent are tightening. Our generated files include an inaudible watermark to mark them as synthetic, helping creators meet emerging disclosure standards. Clear attribution and understanding of licensing, such as the CC BY 4.0 source material, remains essential for commercial use.
- Inaudible watermarking for transparency
- Need for clear synthetic audio disclosure
- Licensing awareness for commercial content
Adoption by content creators
Adoption is driven by the need for consistency and speed. Creators use preset voices to maintain a uniform brand voice across thousands of posts without the cost and complexity of cloning. This approach allows for instant scalability, letting bloggers focus on writing rather than audio production logistics.
- Consistent brand voice across all content
- No waiting for model training
- Cost-effective scaling for high-volume blogs
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Her tone works best for personal essays or lifestyle blogs where connection matters.
23 languages
Any voice speaks any of these languages.
