Frequently asked questions
What does TTS stand for?
TTS stands for Text-to-Speech. Our cloud API converts your input text into an audio file. You send text via a POST request, and receive the synthesized audio in formats like MP3 or WAV. It is a straightforward way to add voice to your mobile app without managing complex infrastructure.
Is the text sent to a server?
Yes, text is sent to our server for synthesis. Input text is used only to produce the audio and is not used for training. Request metadata like time and character count is kept for billing, but the generated audio is returned in the response and not stored.
Can I use the audio offline?
Yes, once you receive the audio file, you can use it offline. The synthesis happens on our server, but the resulting MP3, WAV, or other format files are yours to download and store locally for playback without needing an active internet connection.
How is the audio generated?
Audio is generated by sending a POST request to our API endpoint with your text, voice preset, and format preferences. The server synthesizes the audio using licensed recordings from LibriTTS-R and LibriVox. Each file carries an inaudible watermark marking it as synthetic speech.
Do I need an internet connection?
You need an internet connection to send requests to the API and receive the audio files. However, once the audio file is downloaded to your device, you can play it offline. The synthesis process itself requires a stable connection to our servers.
Is the voice robotic or natural?
The voices are based on licensed recordings from LibriTTS-R and LibriVox, offering natural-sounding speech. We provide eight presets like Aria, Nova, and Felix, covering various tones from warm narrators to energetic promos.
Can I control the speed?
Yes, you can control the speed using the speed parameter in your request. The allowed range is 0.25 to 4.0. This allows you to adjust the playback rate to fit your app’s needs, from slow storytelling to fast-paced instructions.
Do voices change over time?
The voice presets are fixed and do not change over time. They are based on specific licensed recordings. You can choose from eight distinct voices like Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion, and they remain consistent for all your requests.
Is the API secure?
Yes, the API uses Bearer token authentication. You receive a unique key for your account, which you include in the Authorization header. You can issue a new key at any time, and the old one stops immediately, ensuring secure access to your synthesis services.
Can I use this for voiceovers?
Yes, our API is ideal for voiceovers in mobile apps. You can generate audio for instructions, stories, or notifications. With eight voice presets and support for 23 languages, you can find the right tone for your content, from calm instructions to energetic promos.
What is the latency?
Typical speed is roughly real time, about 25 characters of text per second of generation. Short phrases take a few seconds. Requests are processed one after another, so under load, a request may wait in a queue before synthesis begins.
Do I need to buy voices?
No, you do not buy voices separately. The price is $10 per 1 million characters, regardless of the voice used. All eight presets—Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion—are included in the standard per-character rate.
Is there a limit on voice usage?
You are limited by your account balance and rate limits. You can make 60 requests per minute per key. There is no limit on the number of voices you can use, but each request consumes characters from your balance at $0.01 per 1,000 characters.
Can I store the audio files?
Yes, you can store the audio files yourself. The API returns the audio in the response, and we do not store the generated files. You can save them locally or in your cloud storage for offline use in your mobile app.
What is a TTS API?
A Text-to-Speech API is a cloud service that converts written text into audio files. Instead of embedding large voice libraries directly into your mobile app, you send a text payload to a remote server and receive a synthesized audio file in return. This shifts the processing load from the user’s device to our infrastructure, keeping your app lightweight and responsive.
Text to audio conversion
The core function is straightforward: you provide input text, and the system generates corresponding audio. The API handles the complex neural synthesis behind the scenes, delivering clean audio files ready for playback in your app’s interface.
Cloud-based processing
All computation happens on our servers. This means you don’t need to worry about device compatibility or battery drain during synthesis. Your mobile app simply acts as a client, sending requests and playing the resulting audio streams or files.
Standardized endpoints
We use standard HTTP endpoints that integrate easily with most mobile development stacks. The API follows a familiar structure, making it simple to implement using standard libraries without requiring custom binary parsing or complex local engine setups.
How the engine works
Our engine synthesizes speech using neural networks trained on licensed recordings from LibriTTS-R and LibriVox. Each generated file carries an inaudible watermark to mark it as synthetic speech.
Neural network synthesis
Text is processed through neural models to generate natural-sounding audio. The system handles up to 4096 characters per request, processing them sequentially to ensure consistency.
Voice preset selection
You choose from eight licensed presets, such as Aria or Atlas. These voices are based on recordings from public-domain audiobooks, offering distinct tones without custom cloning.
Audio format encoding
The API outputs audio in MP3, WAV, Opus, FLAC, or PCM formats. This flexibility allows you to balance quality and bandwidth for your mobile app.
Request and response flow
When your mobile app needs voice, it sends a standard HTTP POST request to the API endpoint. You include your API key in the authorization header and provide the text, voice preset, and desired audio format in a simple JSON body. This standardized structure works seamlessly with existing OpenAI SDKs by just adjusting the base URL.
Send JSON payload
Construct a JSON object containing the input text (up to 4096 characters), voice preset, and response_format. Include optional parameters like speed (0.25–4.0) or a specific language ISO code if the text script isn't clear. The API validates your key and request structure immediately.
- Use
tts-1ortts-1-hdfor the model - Set
response_formatto mp3, wav, opus, flac, or pcm - Include
Authorization: Bearer your-key-hereheader
Receive audio file
Upon success, the API returns the audio file directly in the response body. You can save this binary data to local storage or play it in your app’s audio player. Failed requests return no charge and clear error codes, ensuring your billing remains accurate.
Handle latency
Generation time roughly matches real-time speech speed (25 characters per second), plus queue time if the server is busy. For short phrases, expect a few seconds; for longer texts, the process takes longer. Implement a loading spinner in your UI to manage user expectations during synthesis.
Why use an API for mobile?
Running synthesis on-device requires significant CPU cycles and battery drain, which can degrade the user experience during other tasks. By offloading the heavy lifting to our cloud API, your mobile app remains responsive and energy-efficient, allowing the device to focus on UI rendering and core logic.
This approach also means you never have to bundle large voice models into your app bundle. Instead, your team delivers a lighter application that updates instantly via the API, ensuring users always hear the latest voice presets without requiring a new app store submission.
Offload processing
Mobile devices vary widely in processing power. Offloading synthesis to our servers ensures consistent audio quality across all devices, from budget phones to high-end flagships, without needing to manage multiple model variants for different architectures.
- Preserves device battery life
- Maintains app responsiveness
- Consistent quality across hardware
Update voices server-side
Voice models can be improved or updated on our end without you releasing a new app version. Your users automatically benefit from better synthesis quality or new voice presets the moment they make a request, keeping your app current with minimal effort.
- Instant voice updates
- No app store delays
- A/B test new voices easily
Reduce app size
Voice models can be several megabytes each. By using a cloud API, you avoid embedding these large assets directly into your APK or IPA file. This keeps your initial download size small, reducing uninstall rates and improving discoverability in app stores.
- Smaller APK/IPA size
- Faster initial download
- More storage for user content
Integration complexity
Integrating our TTS API into your mobile app is straightforward because it relies on standard HTTP calls. You do not need complex native libraries or custom parsers to handle the binary audio data. Simply send a JSON payload with your text and voice selection, and the API returns the audio file directly.
This simplicity reduces the need for native binary parsing logic in your app. You can easily implement caching strategies by storing generated audio files locally on the device or in your app bundle. This approach avoids re-sending identical text requests, saving both API characters and network usage for your users.
Standard HTTP calls
Our API uses standard POST requests with JSON bodies, similar to the OpenAI speech endpoint. This means you can use any common HTTP client library available for iOS or Android. There is no need to learn a proprietary protocol or install heavy SDKs. You just change the base URL and use your API key.
The request structure is simple: specify the model, input text, voice, and desired audio format. This consistency allows your development team to reuse existing networking code for other API calls, reducing the learning curve and speeding up integration timelines for mobile product teams.
- Use any standard HTTP client (OkHttp, URLSession, Retrofit, etc.)
- JSON request body, no complex headers required
- Consistent with widely known API patterns
No native binary parsing
You receive the audio file directly in the response body. This eliminates the need for custom binary parsing logic within your mobile app. Whether you request MP3, WAV, OPUS, FLAC, or PCM, the data is ready to be played or saved immediately.
This reduces the complexity of your audio engine. You can pass the received bytes directly to your app’s media player. This simplicity ensures that your team can focus on UX and core features rather than debugging audio format issues.
- Direct audio file in response body
- No custom binary parsing needed
- Ready for immediate playback or storage
Easy caching strategies
Cache generated audio files locally on the device or in your app bundle to avoid re-sending identical text requests. This is especially useful for static content like menu items, notifications, or accessibility descriptions. By caching, you save API characters and reduce latency for users.
Implementing cache logic is straightforward: check if the audio file for a given text and voice already exists locally before making a request. This strategy also helps manage rate limits (60 requests per minute) by reducing the number of calls needed for repeated content.
- Store audio locally on device or in app bundle
- Avoid re-sending identical text requests
- Save API characters and reduce latency
Voice customization options
Voice customization happens at request time, allowing your mobile app to adapt audio output without code changes. You can adjust speed, select languages, and choose from eight distinct voice presets to match your app’s tone.
These options are passed directly in the JSON payload, giving your product team fine-grained control over the listening experience.
Speed adjustment
Adjust playback speed between 0.25x and 4.0x. Faster speeds help users consume content quickly, while slower rates improve clarity for complex information or accessibility needs.
The engine handles real-time speed adjustments without affecting audio quality or requiring post-processing.
- Range: 0.25 to 4.0
- Applies during synthesis
- No extra cost
Language selection
Explicitly set the ISO language code for accurate pronunciation. While the API guesses language from script, setting it explicitly ensures correct accents for non-English Latin text.
Supports 23 languages, including Arabic, Chinese, Hindi, Japanese, and Spanish, among others. Any voice can speak any supported language.
Voices do not change based on language; you select the voice independently.
- 23 supported languages
- Explicit code recommended
- Language guessing fallback
Voice preset limits
Choose from eight presets: four female (Aria, Nova, Luna, Sage) and four male (Felix, Milo, Atlas, Orion). Each has a distinct tone, from upbeat to documentary.
You cannot create custom voices or clone existing ones. Selection is limited to these eight licensed recordings.
Each voice carries an inaudible synthetic watermark.
- 8 presets total
- 4 female, 4 male
- No custom cloning
- Watermarked audio
Cost implications
Understanding the billing model helps you forecast expenses accurately for your mobile app’s audio features. VocalPipe charges based on the actual characters processed, including spaces and punctuation, ensuring you pay only for what is generated.
Per-character billing
The rate is $10 per million characters, which equals $0.01 per 1,000 characters. This straightforward metric makes it easy to estimate costs based on your expected daily usage volume without hidden fees.
Network usage costs
Since audio files are returned directly in the response, your app incurs standard data transfer costs. Using efficient formats like mp3 or opus helps minimize the bandwidth required per request compared to uncompressed audio.
Caching to save money
If your app frequently plays the same audio, such as static notifications or menu sounds, you can cache the generated files locally. This avoids sending repeated requests for identical text, reducing both API calls and data usage.

Latency considerations
Understanding latency helps you set the right expectations for your mobile users. Our cloud API processes requests sequentially, meaning short phrases take a few seconds while longer texts generate at roughly real-time speed (25 characters per second).
Knowing these limits allows you to design intuitive loading states or background processing flows that keep your app feeling responsive without misleading users about generation time.
Real-time vs pre-generated
Unlike pre-generated audio files stored in your app bundle, our API generates audio on demand. This ensures voice updates are instant but introduces variable generation time based on input length and current server load.
For short phrases, expect a few seconds of processing. For longer texts, the generation speed averages near real-time, meaning a 60-second audio file takes approximately 60 seconds to create.
- Short texts: a few seconds processing
- Long texts: ~25 characters per second
- No pre-stocked audio files
User experience impact
Latency directly affects perceived app performance. If a user requests a notification sound or a short UI feedback voice, the delay must be minimal to avoid frustration.
Design your UI to handle these few seconds gracefully. Use skeleton loaders or clear "generating" indicators so users understand the process is active and not stuck.
For longer content like audiobook chapters, consider generating content in the background during user interactions to hide the latency entirely.
Queue waiting times
Requests are processed one after another, so under heavy load, your request may wait in a queue before synthesis begins.
This queuing is normal and transparent; you will receive a response once processing is complete, but the total time may exceed the generation duration if many users are active simultaneously.
Rate limits (60 requests/minute) also influence how quickly your requests enter the processing pipeline, ensuring stable audio quality even during spikes.
- Sequential processing under load
- Queuing adds to total latency
- Rate limits affect entry speed
Data privacy
When your app sends text to our cloud API, the input is transmitted securely to our servers for synthesis. This data travels over standard HTTPS connections, ensuring that your user’s voice content remains private during the brief window of processing. We treat this data as transient, using it solely for the immediate generation of the audio file.
Once the synthesis is complete, the resulting audio is returned directly to your client or server. We do not retain the generated audio files after they are delivered, meaning your users’ spoken content does not accumulate in our long-term storage. This minimizes the risk of data exposure for sensitive app features.
Text sent to cloud
Your application transmits the text payload to our endpoint. The text is processed in real-time to generate the corresponding audio waveform. This transmission is necessary for our neural engine to apply the selected voice preset and language rules.
Audio not stored
The generated audio file is returned in the API response and is not saved on our servers. Only request metadata, such as timestamp and character count, is retained for billing purposes. This ensures that your users’ spoken content remains ephemeral and is not archived in our infrastructure.
No training usage
Unlike some consumer-grade services, we do not use your input text or generated audio to train our models. The data you send is strictly used for the single purpose of synthesizing the requested audio file. Your proprietary content remains yours, with no secondary usage rights granted to us.
Scalability
Our API handles concurrent requests by processing them sequentially. Under heavy load, requests may wait in a queue, but throughput remains consistent. This ensures your mobile app receives audio without crashing, even during spikes.
Handle peak loads
Requests are processed one after another. If many users request audio simultaneously, your requests may wait in a queue. This linear processing prevents system overload while maintaining predictable performance for your app.
Rate limit management
You have a limit of 60 requests per minute per key. New accounts start with 10 requests per minute. If you exceed this, requests may fail. You can issue a new API key to reset limits immediately, as the old key stops working at once.
- Limit: 60 requests/minute (standard)
- New accounts: 10 requests/minute
- Key rotation resets limits instantly
Global content delivery
Our API serves from cloud infrastructure. We do not guarantee specific uptime or data-center locations. For large-scale mobile apps, implement client-side caching to reduce API calls and minimize latency for users worldwide.
Comparison with SDKs
We provide a standard REST API rather than a proprietary mobile SDK. This approach keeps your app bundle small and avoids platform-specific dependencies, relying instead on universal HTTP calls that any modern mobile framework can handle.
API vs local engine
Unlike local engines that consume significant device storage and CPU cycles, our cloud API offloads synthesis to our servers. This preserves battery life and processing power for your app’s core features, delivering high-quality audio without the overhead of running neural networks on the user’s device.
- Zero local model storage required
- Consistent audio quality across all devices
- No background process management needed
Update frequency
When we improve our voice models or add new formats, mobile users receive the updates instantly without waiting for an App Store or Play Store review. Your app only needs to send text; the latest synthesis engine handles the rest, ensuring your users always hear the available quality.
Device compatibility
Our API works uniformly across iOS, Android, and any hybrid framework since it relies on standard JSON requests. You do not need to maintain separate native implementations for different architectures. This simplifies your development workflow and ensures consistent behavior whether the user is on a low-end phone or a flagship tablet.
Use cases in mobile apps
Mobile apps leverage our TTS API to convert text into audio files on the fly, enabling rich media experiences without heavy local infrastructure. By offloading synthesis to our cloud, your team maintains control over UX while delivering consistent voice quality across devices.
Accessibility features
Integrate spoken feedback for visually impaired users by sending screen text to our endpoint. Since voices support 23 languages, you can automatically detect the user's language or let them choose a preset like Aria or Felix for clear, narrator-style audio.
Audio books
Generate chapter summaries or full audiobooks by splitting long texts into 4096-character chunks. Use the tts-1-hd model for higher fidelity, then join the resulting audio files. This approach keeps your app bundle small while offering a library of voices.
Notifications
Convert push notification text into short audio clips using the standard mp3 or opus format. With speeds adjustable from 0.25 to 4.0, users can listen to urgent updates at a comfortable pace without reading a screen.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Our mobile apps must deliver seamless user experiences that keep product teams engaged and satisfied daily.
23 languages
Any voice speaks any of these languages.
