Frequently asked questions
What is TTS?
Text-to-speech (TTS) is a cloud API that converts written text into audio files. You send a POST request with your text, and the service returns synthesized speech. It is designed for developers to embed voice capabilities directly into their smart home devices and IoT applications without managing complex audio infrastructure.
How does TTS work in IoT?
Your device sends text via HTTP to our endpoint. We process the request and return an audio file. For long texts, split them into 4096-character parts and join the resulting audio files. This simple workflow integrates easily with existing smart home protocols, allowing your devices to speak clearly and reliably.
What are the benefits?
Our API offers clear, multilingual output optimized for embedded systems. You get predictable pricing at $0.01 per 1,000 characters, with failed requests being free. The solution supports 23 languages and multiple audio formats, ensuring your smart devices can communicate effectively with users across different regions and technical environments.
What are the use cases?
Use our API for voice assistants, smart speaker notifications, and interactive product instructions. It is ideal for smart home devices that need to read out status updates, answer queries, or provide audio feedback. The lightweight integration makes it suitable for resource-constrained IoT hardware requiring reliable synthetic speech.
What are the technical requirements?
You need an internet connection to send POST requests to our API. The input text must be under 4096 characters per request. You will need an API key for authentication. Our SDKs work by changing the base URL, allowing you to use standard HTTP libraries or the OpenAI SDK for integration.
How do I choose a voice?
We offer eight presets, including female voices like Aria and Nova, and male voices like Felix and Milo. Each voice has a distinct tone, from warm narrators to energetic promo styles. You can select any voice for any language, ensuring the audio matches your device’s brand personality and user experience goals.
Which languages are supported?
We support 23 languages, including English, Spanish, German, French, Chinese, Japanese, and Arabic. Any voice can speak any of these languages. If you do not set the language parameter, we guess it from the script, but setting it explicitly ensures accuracy for non-English Latin-script text.
What audio formats are best?
You can choose from mp3, wav, opus, flac, or pcm. MP3 is efficient for storage and transfer, while WAV offers uncompressed quality. Opus and FLAC provide high efficiency or lossless compression. PCM is raw data. Choose based on your device’s storage constraints and playback capabilities for optimal performance.
Is latency an issue?
Processing takes roughly real-time speed, about 25 characters per second. Short phrases may take a few seconds. Requests are processed sequentially, so under heavy load, a request might wait in a queue. This is generally acceptable for most smart home interactions where immediate millisecond response is not critical.
Is data private?
Your input text is used only to produce the audio and is not used for training. Generated audio is returned in the response and not stored. We keep request metadata like time and status for billing purposes. This ensures your voice data remains private and secure during processing.
How scalable is the API?
The API handles requests via a cloud endpoint. You have a rate limit of 60 requests per minute per key. For higher volumes, you can issue a new key (the old one stops instantly) or manage multiple keys. The system processes requests sequentially, so plan your architecture to handle queuing during peak usage periods.
Can I use it offline?
No, this is a cloud API. You must send text to our endpoint to receive audio. There is no offline mode or local engine. Your device needs an active internet connection to make POST requests and receive the synthesized audio files for playback.
Do I need a speaker?
Yes, to hear the output, your device must have a speaker or audio output interface. The API returns an audio file, which your device’s hardware plays. Without a speaker, you can still receive the audio data for storage or further processing, but users will not hear the speech.
How do I test it?
Sign up on our site to get a free trial with $0.50 credit. Use the provided curl example or OpenAI SDK to send a test request. Check the /docs/ page for endpoint details. You can verify the audio quality and language support before integrating the API into your production smart devices.
Definition of TTS in IoT
Text-to-speech (TTS) in IoT transforms written text into audible speech, enabling smart devices to communicate naturally with users. This cloud-based voice generation ensures consistent audio quality across various hardware, from smart speakers to home security systems.
By integrating our API, product teams can embed voice capabilities directly into their devices. This approach simplifies development, allowing teams to focus on hardware and user experience while relying on a reliable, scalable backend for audio synthesis.
Text to audio conversion
TTS converts digital text into human-like speech, allowing devices to read notifications, instructions, or responses aloud. This capability is essential for creating intuitive interfaces, especially in environments where visual attention is limited.
- Converts any text input into spoken audio
- Supports natural intonation and pacing
- Ideal for hands-free device interactions
Cloud-based voice generation
Our API handles the heavy lifting of voice synthesis in the cloud. Devices send text via API requests and receive audio files, reducing the need for powerful processors or large memory banks on the hardware itself.
- Offloads processing to the cloud
- Saves device battery and resources
- Ensures consistent audio quality
API integration for devices
Integration is straightforward using standard HTTP requests. By pointing the OpenAI-compatible SDK to our endpoint, teams can quickly add voice features to existing smart home or IoT products without rebuilding core infrastructure.
- Simple POST requests for speech synthesis
- Compatible with standard IoT protocols
- Easy to embed in firmware or apps
How TTS Works in Smart Devices
Our cloud API handles the heavy lifting: send text via a standard HTTP POST, and we return a synthesized audio file. This simple workflow lets embedded systems focus on execution rather than audio processing.
The process begins with a JSON request specifying the model, voice, and input text. Our servers process the request sequentially, ensuring consistent output quality for every device connected to your network.
API request structure
Send a POST request to https://api.playll88.com/v1/audio/speech with a JSON body. Include your Authorization: Bearer key, the text input (up to 4096 characters), and your chosen voice preset. You can also specify response_format (mp3, wav, opus, flac, pcm) and an optional language code.
The endpoint accepts parameters identical to the OpenAI speech API, allowing you to use existing OpenAI SDKs by simply changing the base_url. This reduces integration time for product teams already familiar with standard TTS workflows.
- Endpoint:
POST https://api.playll88.com/v1/audio/speech - Auth:
Authorization: Bearer tts-... - Input: JSON body with
input,voice,model,response_format
Audio file generation
Once the request is received, our servers synthesize the audio using licensed voice models. The generation typically takes a few seconds for short phrases, scaling to roughly real-time speed for longer texts. If synthesis fails, you are not charged.
Each generated file is returned directly in the HTTP response. You can also use the GET /v1/voices endpoint to preview available voices and languages before committing to a specific configuration for your device firmware.
Playback on device
Your smart device receives the audio file and plays it through the local speaker. Since we do not offer streaming, you may need to buffer the entire file before playback begins, depending on your network conditions.
For devices with limited memory, choose efficient formats like MP3 or OPUS. For high-fidelity requirements, WAV or FLAC are available. The audio includes an inaudible watermark marking it as synthetic speech, useful for transparency in consumer applications.
Benefits for Product Teams
The SmartVoice API streamlines development by allowing product teams to integrate voice features without building complex audio pipelines. With a simple POST request, you receive ready-to-play audio files, accelerating the transition from concept to functional prototype.
This efficiency is particularly valuable for IoT projects where hardware constraints vary. By offloading synthesis to the cloud, teams can focus on device logic and user experience rather than managing local audio engines or codec conversions.
Faster prototyping
Sending text via API yields audio files instantly, enabling rapid iteration on voice interactions. Teams can test different phrasing and timing without reprogramming the device’s local memory or updating firmware for every minor change.
The $0.50 free trial provides immediate access to 50,000 characters, allowing developers to validate voice integration before committing to a paid plan. This low barrier to entry reduces risk during early development stages.
- No local audio engine required
- Immediate feedback loop for voice UX
- Low-cost trial for validation
Multilingual support
With 23 languages available, teams can deploy a single API key for global devices. The API automatically detects language from script or accepts explicit ISO codes, ensuring accurate pronunciation without maintaining separate language models.
This capability simplifies internationalization. Product teams can localize voice assistants and alerts for markets like Japan, Germany, or Brazil using the same integration code, reducing translation and engineering overhead.
- 23 languages supported
- Automatic or manual language setting
- Single integration for global markets
Consistent voice output
Using one of eight preset voices ensures uniform audio quality across all devices. Whether using the calm 'sage' for instructions or the energetic 'felix' for promos, the output remains consistent regardless of device hardware.
Each audio file includes an inaudible watermark, helping teams verify synthetic origin if needed. This consistency builds user trust in the device’s voice identity, crucial for brand recognition in smart home ecosystems.
- 8 distinct voice presets
- Uniform quality across devices
- Inaudible watermark for verification
Common Use Cases
Smart devices rely on clear audio cues to communicate effectively with users. Our API enables product teams to generate natural-sounding voice responses directly from text commands, ensuring seamless interaction without requiring complex local speech synthesis engines.
Voice assistant responses
When a user asks a smart speaker a question, the API converts the answer text into an audio file via a simple POST request. Teams can choose neutral voices like Atlas for news or Felix for promotional content, ensuring the tone matches the brand.
- Instant conversion of query answers to audio
- Consistent voice identity across all devices
- Supports 23 languages for global markets
Security alerts
For security systems, clarity is critical. The API generates clear audio alerts for motion detection or door openings. Using a calm voice like Sage ensures the alert is heard but not overly alarming, providing precise instructions for smart home integration.
- Clear audio for critical notifications
- Low latency for real-time alerts
- Reliable delivery over standard networks
Device status updates
Smart thermostats or robot vacuums can announce their status using generated speech. Developers send short text strings via the API to produce MP3 or WAV files, which the device plays immediately. This removes the need for pre-recorded audio files for every language.
- Dynamic status announcements
- Multilingual support without extra storage
- Easy integration with existing IoT frameworks
Technical Requirements
Smart devices require a stable network connection to send text payloads to the API and receive audio files. Since requests are processed sequentially, ensure your IoT infrastructure can handle brief queueing during peak usage.
Hardware speakers must support the selected output format. While most devices play MP3 or WAV natively, optimized formats like OPUS reduce bandwidth usage for devices with limited memory or slow connections, ensuring smoother playback of generated voice responses.
Network connectivity
A reliable internet connection is essential for sending text requests and receiving the resulting audio file. Devices should handle brief latency gracefully, as requests may queue during high load.
Ensure your device firmware can manage transient network interruptions, as the API returns a complete audio file rather than a continuous stream.
- Stable Wi-Fi or cellular connection required
- Handle temporary latency without audio glitches
Speaker hardware
Your device’s audio output must support the chosen response format, such as MP3, WAV, or OPUS. Lower bitrate formats like OPUS are ideal for constrained IoT devices to save bandwidth.
Ensure the speaker driver can reproduce the full frequency range of the selected voice preset for clear, intelligible speech in home environments.
- Support for MP3, WAV, OPUS, FLAC, or PCM
- OPUS recommended for bandwidth-constrained devices
API integration
Integrate using standard HTTP POST requests to https://api.playll88.com/v1/audio/speech. You can use the OpenAI SDKs by simply changing the base_url, or send raw JSON with your Authorization: Bearer key.
Each device needs a valid API key. Rotate keys periodically via the /get-api-key/ dashboard to maintain security without disrupting ongoing operations.
- Use POST
https://api.playll88.com/v1/audio/speech - OpenAI SDKs compatible via
base_urlchange - One active key per account; rotate via dashboard
Voice Selection for IoT
Choosing the right voice preset ensures your smart device’s audio feedback is both clear and context-appropriate. The API provides eight distinct voices, allowing product teams to match the tone of the interaction, whether it’s a simple status update or a complex instruction.
Clear and understandable voices
Use Aria or Sage for clear narration. Aria offers a warm, clear tone ideal for general announcements, while Sage provides a calm, instructional style perfect for step-by-step guides.
- Aria: Warm, clear narrator
- Sage: Calm, instructional
Neutral tones for instructions
Atlas delivers a neutral, news-style voice suitable for straightforward commands and alerts. Milo offers a friendly, conversational tone that feels less robotic for everyday interactions.
- Atlas: Neutral, news-style
- Milo: Friendly, conversational
Energetic tones for engagement
For promotional features or active engagement, use Felix’s energetic promo style or Nova’s bright, upbeat delivery. Orion’s deep, documentary tone works well for immersive experiences.
- Felix: Energetic, promo
- Nova: Bright, upbeat
- Orion: Deep, documentary

Language Support
Our API supports 23 languages, enabling smart home devices to communicate naturally with users globally. From Arabic to Swahili, each language is handled by the same robust engine, ensuring consistent quality across your connected ecosystem. This multilingual capability allows product teams to deploy voice assistants in diverse markets without managing separate voice models for each region.
Devices can operate seamlessly in mixed-language environments, adapting to user preferences or regional settings. Whether your device is in Tokyo, Berlin, or São Paulo, the TTS service delivers clear audio output tailored to the local language, enhancing user experience and accessibility.
23 languages available
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. Any voice can speak any of these languages, giving you flexibility in voice selection without language restrictions.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Automatic language detection
If you do not set the language field, the API guesses the language from the script. Latin-script text defaults to English, which works well for most Western smart home contexts. This simplifies integration for devices that primarily communicate in English.
Manual language setting
For non-English Latin-script text, you should set the language parameter explicitly to ensure accurate pronunciation. Pass the ISO code to guide the synthesis engine. This prevents misinterpretation and ensures the audio matches the expected language for your device's audience.
Audio Formats for Devices
Smart devices often have strict constraints on bandwidth and processing power. Choosing the right audio format ensures your voice assistant responds quickly without draining battery or consuming excessive data plans.
MP3 compatibility
MP3 is the standard for broad compatibility. It offers a good balance between file size and sound quality, making it ideal for most IoT speakers and voice assistants.
WAV quality
WAV provides uncompressed, high-fidelity audio. Use this format when audio quality is prioritized over file size, such as in high-end smart home hubs or when precise audio analysis is required downstream.
OPUS efficiency
OPUS delivers superior compression, reducing data usage significantly. This is crucial for devices with limited connectivity or those operating on metered networks, ensuring faster playback initiation and lower operational costs.
Latency Considerations
Synthesis latency depends on text length and queue status. Long texts generate at roughly real-time speed (25 characters per second). Short phrases take a few seconds. Requests process sequentially, so high load may cause queuing delays.
Network round-trip time adds overhead. Ensure stable connectivity for consistent playback timing.
Real-time generation
Generation speed matches real-time audio duration for long texts. Short inputs complete in seconds. Sequential processing means concurrent requests may queue.
Use the 4096-character limit efficiently. Split longer scripts to manage queue wait times predictably.
- Approx. 25 characters generated per second of audio.
- Sequential processing; concurrent requests may queue.
- Short phrases take a few seconds to synthesize.
Network delay factors
Latency includes API request time and audio download. Ensure reliable internet connectivity on the device.
Minimize payload size by choosing efficient formats like OPUS or MP3 over WAV or FLAC where quality permits.
- Stable network connection reduces jitter.
- Format choice impacts download time.
- Sequential API processing adds queue time under load.
Optimization techniques
Pre-fetch audio for common phrases. Cache responses locally on the device to avoid repeated API calls.
Split long texts into chunks under 4096 characters. Join audio files locally to maintain seamless playback.
- Cache common responses on-device.
- Split text into <4096 char chunks.
- Use OPUS for smaller payloads.
Data Privacy
Our API is designed with privacy in mind for smart devices. When you send text to our cloud, we process it solely to generate the audio response. Once the audio file is created and returned to your device, the original text input is not retained on our servers.
Similarly, the generated audio file is not stored after delivery. This ensures that voice data remains transient, reducing the risk of long-term data accumulation from your connected devices.
Text not stored
Input text is used exclusively for the immediate synthesis of audio. After the audio file is generated and returned in the response, the text is not kept on our servers.
- Text is processed only for audio generation
- No long-term storage of input text
Audio not stored
Generated audio files are returned directly in the API response and are not stored by our service. This ensures that voice outputs remain transient and are not archived for later use or training.
- Audio is returned in the response
- No storage of generated audio files
Metadata kept for billing
We retain request metadata, including timestamp, character count, and status, strictly for billing purposes. This data helps us accurately calculate usage charges based on your character count.
- Timestamps kept for billing
- Character counts recorded
- Status codes logged
Scalability
Our API scales to support thousands of connected devices without requiring complex infrastructure. Each request consumes characters from your balance, making cost predictable regardless of device count. The system processes requests sequentially, ensuring stability even during peak usage times.
Handle multiple devices
A single API key can serve numerous smart home hubs simultaneously. Since audio files are returned directly in the response and not stored, you avoid bandwidth bottlenecks from server-side storage. This design keeps device communication lightweight and fast.
Rate limits management
Each key allows 60 requests per minute once topped up. For smaller deployments, this limit is rarely reached. If you anticipate high traffic, monitor your usage dashboard to stay within the 10 requests/minute baseline before top-up.
Key rotation strategies
You can issue a new key at any time via the dashboard. The old key stops instantly, allowing seamless transitions during security audits or device fleet updates without downtime. This ensures continuous voice service for your users.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Connect devices seamlessly to deliver intuitive, responsive living experiences.
23 languages
Any voice speaks any of these languages.
