Frequently asked questions
Which TTS API is best for IoT?
HomeEcho provides deterministic audio output for smart home developers. Our API ensures consistent results for voice assistants and device notifications without external dependencies. We support 23 languages and eight voice presets, handling scale efficiently for IoT applications requiring reliable synthetic speech generation.
Is HomeEcho better than OpenAI?
HomeEcho is not affiliated with OpenAI. Our API accepts the same request shape as OpenAI’s speech endpoint, allowing you to use official OpenAI SDKs by changing the base_url. We offer identical models like tts-1 and tts-1-hd at competitive rates without requiring OpenAI accounts.
Does HomeEcho support SSML?
No, HomeEcho does not support SSML. The API accepts plain text input up to 4096 characters. It does not offer phoneme dictionaries or custom pronunciation lexicons. You control pacing via the speed parameter, which ranges from 0.25 to 4.0, without markup tags.
Which voice is best for notifications?
For notifications, we recommend atlas for neutral news-style alerts or felix for energetic promo-style notifications. Atlas provides a clear, neutral tone suitable for general updates, while felix offers an energetic delivery. Both are available in all supported languages and formats for consistent device audio.
How does HomeEcho compare to Google?
HomeEcho is not affiliated with Google. We provide a standard REST API compatible with OpenAI SDKs. Unlike Google’s varied ecosystem, we offer eight specific voice presets and direct USDT payments via TRON. Our pricing is transparent at $10 per million characters with no hidden tiers.
Is Amazon Polly more expensive?
HomeEcho charges $10 per million characters. Failed requests are free, with charges returned in full. We do not offer monthly plans or subscriptions. Payment is via USDT on TRON. Pricing is straightforward, with no per-request tier variations or complex volume discounts compared to other providers.
Can I use HomeEcho for smart speakers?
Yes, HomeEcho is designed for smart home and IoT developers. You can generate audio files via POST requests and play them on devices. The API supports formats like mp3 and wav, ensuring compatibility with standard smart speaker audio players for voice assistants and notifications.
What is the best format for IoT?
We support mp3, wav, opus, flac, and pcm. Mp3 is efficient for bandwidth-constrained IoT devices. Wav offers uncompressed quality. Opus provides high compression with good quality. Choose based on your device’s decoding capabilities and storage constraints, as all formats are generated at the same speed.
Does HomeEcho have a free tier?
Accounts created with Google sign-in receive a $0.50 trial credit valid for 7 days, covering about 50,000 characters. Email-based accounts start with zero credit. There is no permanent free tier; you must top up via USDT to continue using the API after credits are exhausted.
Which language has the best quality?
All 23 supported languages use the same engine and voice presets. Quality is consistent across Arabic, Chinese, English, and others. Voices are based on licensed recordings. There is no single best language; each voice preset adapts to the linguistic nuances of the selected language automatically.
Is the API easy to integrate?
Yes, the API uses a standard JSON body similar to OpenAI’s speech endpoint. You can use official OpenAI SDKs by changing the base_url. Send a POST request with your key, model, input text, and voice. The response returns the audio file directly, simplifying integration for developers.
Does HomeEcho store my data?
Input text is used only to produce audio and is not used for training. Generated audio is returned in the response and not stored. We keep request metadata like time, characters, and status for billing purposes. Your raw text content is not retained after synthesis.
Can I clone voices with HomeEcho?
No, HomeEcho does not offer custom voice cloning. You cannot upload your own voice or design custom voices. We provide eight fixed presets: aria, nova, luna, sage, felix, milo, atlas, and orion. Each voice is based on licensed recordings from LibriTTS-R and LibriVox public-domain sources.
What is the latency like?
Typical speed is roughly real time, generating about 25 characters per second of audio. Short phrases take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue. Latency depends on server load and text length, not fixed SLAs.
Is USDT the only payment method?
Yes, top-ups are accepted only via USDT on the TRON network (TRC-20). We do not accept credit cards, PayPal, or subscriptions. The minimum top-up is $5. Credits are added automatically within minutes of transfer. No invoices or VAT are generated for these transactions.
Are there hidden fees in HomeEcho?
No, pricing is $10 per million characters. Failed requests are free. There are no monthly plans, seats, or minimum spends. Top-ups via USDT are the only cost. You pay only for successful character generation. No hidden fees or surprise charges apply to your account.
Evaluation Criteria
Selecting a text-to-speech API for smart home and IoT deployments requires evaluating three core dimensions: pricing structure, audio fidelity, and integration flexibility. IoT developers need predictable costs for device notifications and consistent voice output for voice assistants without external dependencies.
This guide compares HomeEcho against major providers based on these criteria, focusing on deterministic results and scale capabilities suitable for indie developers building connected home ecosystems.
Pricing Model
Pricing structures vary significantly across providers. HomeEcho charges $0.01 per 1,000 characters with no minimum spend or monthly plans. Failed requests incur no charge. Payment is processed via USDT on the TRON network, eliminating traditional subscription overhead.
Competitors often utilize tiered enterprise contracts or per-minute billing models that complicate budget forecasting for small-scale IoT deployments.
- $0.01 per 1,000 characters
- No monthly plans or seats
- Failed requests are free
Voice Quality
Voice quality directly impacts user experience in smart home environments. HomeEcho offers 8 distinct voice presets, including Aria for narration and Atlas for news, all derived from licensed recordings. Every file includes an inaudible synthetic watermark.
Unlike generic TTS engines, these presets are optimized for clarity and consistency, ensuring deterministic audio output critical for voice assistant interactions and device alerts.
- 8 preset voices
- Licensed source recordings
- Inaudible synthetic watermark
API Flexibility
HomeEcho uses a standard JSON API compatible with the OpenAI SDK by adjusting the base URL. This allows developers to leverage existing libraries without custom integrations. The endpoint supports multiple audio formats including MP3, WAV, and Opus.
The API requires no streaming or WebSocket connections for standard synthesis, simplifying network requirements for resource-constrained IoT devices. Authentication is handled via a single Bearer token.
- OpenAI SDK compatible
- Multiple audio formats
- Simple Bearer token auth
HomeEcho Overview
HomeEcho provides reliable text-to-speech synthesis for smart home and IoT developers who need consistent audio output without external dependencies. Our API processes requests at scale, ensuring reliable generation for voice assistants and device notifications. The system supports 23 languages, allowing any voice preset to speak any supported language, with automatic language detection for Latin-script text or explicit specification via the language ISO code.
8 Voice Presets
We provide eight voice presets based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobook readings. Each generated audio file carries an inaudible watermark marking it as synthetic speech. You can select from female voices like Aria for clear narration or Luna for storytelling, and male voices like Felix for energetic promos or Orion for deep documentary styles.
- Aria: warm, clear narrator
- Nova: bright, upbeat
- Luna: soft, storytelling
- Sage: calm, instructional
- Felix: energetic, promo
- Milo: friendly, conversational
23 Languages
HomeEcho supports 23 languages including Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. Voices are not language-specific; any voice can speak any of these languages. For non-English Latin-script text, explicitly set the language parameter to avoid default English detection.
- Arabic
- Chinese
- Danish
- Dutch
- English
- Finnish
USDT Payment
Payments are processed via USDT on the TRON network (TRC-20) only, with a minimum top-up of $5. Credits are added automatically within minutes of transfer. There are no monthly plans, no seats, and no minimum spend. Balances do not expire, and failed requests are refunded in full. We do not accept cards, PayPal, or issue VAT invoices.
OpenAI Comparison
HomeEcho utilizes the same engine as OpenAI’s TTS-1 model, delivering deterministic audio synthesis for smart home applications. By accepting identical input parameters, our API ensures consistent voice output without requiring complex configuration adjustments during device integration.
Developers can leverage official OpenAI SDKs by simply redirecting the base URL to https://api.playbk3.com/v1. This compatibility reduces integration overhead, allowing IoT engineers to deploy voice assistants using familiar code structures while maintaining precise control over audio generation.
TTS-1 Model
Our infrastructure runs the same engine as OpenAI’s TTS-1, ensuring consistent voice synthesis for IoT devices.
- Same engine as OpenAI TTS-1
- Deterministic audio synthesis
- Uniform voice output
SDK Compatibility
The API mirrors the request shape of the OpenAI speech endpoint, enabling direct use of official OpenAI SDKs. Developers only need to update the base URL to https://api.playbk3.com/v1, significantly reducing integration time for existing codebases.
- OpenAI SDK compatible
- Simple base URL change
- Reduced integration overhead
Price Difference
HomeEcho offers a flat rate of $10 per million characters, matching OpenAI’s pricing for the tts-1 model. This predictable cost structure supports budget planning for high-volume IoT deployments without unexpected fees or tiered pricing complexities.
- $10 per million characters
- Matches OpenAI tts-1 pricing
- Predictable cost structure
Google Cloud TTS
Google Cloud TTS offers neural voices with advanced intonation, providing a more natural baseline for smart home devices compared to standard synthesis. Its infrastructure supports high availability, making it suitable for enterprise-grade IoT deployments requiring consistent voice output across global regions.
Neural Voices
The platform utilizes deep learning models to generate highly natural speech. These neural voices adapt to context, reducing the robotic tone often found in basic TTS solutions used in consumer electronics.
- Deep learning-based voice synthesis
- Context-aware intonation
- High naturalness for voice assistants
SSML Support
Google Cloud TTS provides extensive SSML (Speech Synthesis Markup Language) support, allowing developers to control pauses, pitch, and speaking rate with precision. This is critical for IoT applications requiring specific audio cues or emphasis in device notifications.
- Full SSML compliance
- Precise timing control
- Customizable pitch and rate
Enterprise Features
The service includes robust enterprise features such as dedicated support channels, SLA guarantees, and extensive data residency options. These features cater to large-scale IoT deployments where reliability and compliance are paramount, though they often come with higher complexity and cost compared to simpler API solutions.
- SLA guarantees
- Dedicated support channels
- Data residency options
Amazon Polly
Amazon Polly remains a dominant choice for IoT developers already embedded in the AWS ecosystem, offering neural text-to-speech synthesis with low latency. Its infrastructure scales automatically, making it suitable for high-volume smart home notifications and voice assistant responses without manual server management.
Lifelike Voices
Polly provides a range of neural voices optimized for natural-sounding output in enterprise environments. While the voice library is extensive, it lacks the specific preset naming convention of newer competitors, requiring developers to map standard voice IDs to their application logic.
- Support for multiple languages and regional accents
- Neural TTS engine for enhanced naturalness
- Standard and neural voice options
AWS Integration
The primary advantage is seamless integration with existing AWS services, allowing direct storage of generated audio in S3 or integration with IoT Core rules. This reduces the complexity of building custom storage and delivery pipelines for smart device notifications.
- Direct S3 integration for audio storage
- IoT Core rule integration
- Existing AWS IAM role compatibility
Pay-Per-Character
Polly charges based on characters processed, similar to our per-character model, but pricing varies by voice type and language. Unlike our flat $10 per million characters rate, Polly’s costs can escalate quickly with high-volume IoT deployments, especially when using premium neural voices.
- Cost varies by voice type
- Price increases with premium neural voices
- No flat-rate bulk discount structure
Voice Selection Analysis
Selecting the correct voice preset is critical for deterministic audio output in smart home environments. HomeEcho provides eight distinct presets, each optimized for specific use cases, ensuring consistent tone across device notifications and voice assistant responses.
Aria vs Standard Female
Aria offers a warm, clear narrator style suitable for general instructions. Unlike standard female voices that may sound robotic, Aria provides a more natural listening experience for device alerts.
- Warm, clear narrator tone
- Optimized for general instructions
- Natural listening experience
Atlas for News
Atlas delivers a neutral, news-style voice ideal for weather updates or system status reports. Its balanced delivery ensures clarity without emotional bias, making it reliable for informational broadcasts.
- Neutral, news-style delivery
- Ideal for weather updates
- Balanced, unbiased tone
Orion for Documentaries
Orion features a deep, documentary-style voice perfect for longer audio content or guided tours. Its rich timbre adds depth to pre-recorded or generated audio narratives in IoT applications.
- Deep, documentary-style voice
- Suitable for longer content
- Rich, deep timbre
Language Support Comparison
Language coverage determines whether your smart home or IoT device can communicate effectively with users across different regions. HomeEcho supports 23 languages, allowing any of the 8 voice presets to speak any supported language, ensuring consistent audio output regardless of the target locale.
Arabic and Chinese
HomeEcho supports both Arabic and Chinese, enabling voice assistants to serve non-Latin script users without external translation dependencies. Unlike some competitors, our voices handle these languages natively, though you must explicitly set the language field for non-English Latin text to ensure correct pronunciation.
European Languages
We support a broad range of European languages, including German, French, Spanish, Italian, Dutch, Polish, Swedish, Norwegian, Danish, Finnish, Russian, Greek, and Turkish. This allows IoT developers to deploy devices in diverse European markets with a single API integration, maintaining uniform voice characteristics across all supported locales.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Excluded Languages
HomeEcho does not support Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, or other languages outside the listed 23. If your IoT device targets a region using an excluded language, you must integrate a separate translation or TTS service, as our API will not generate audio for those scripts.

Technical Integration
HomeEcho utilizes a standard RESTful architecture, mirroring the OpenAI speech endpoint structure. This ensures compatibility with existing HTTP clients without requiring specialized protocols. Developers send text payloads via POST requests to receive audio files directly in the response body.
REST vs gRPC
The API relies exclusively on HTTP POST requests for audio generation. There is no support for gRPC, WebSocket streaming, or real-time partial audio delivery. Clients must manage request splitting for texts exceeding 4096 characters and concatenate the resulting audio files manually.
SDK Availability
HomeEcho does not provide a proprietary SDK. Instead, it adopts the OpenAI SDK request shape. Developers can use the official OpenAI SDKs for Python, Node.js, or Go by simply changing the base_url to https://api.playbk3.com/v1 and supplying the API key.
Authentication Methods
Access is controlled via a single active Bearer token passed in the Authorization header. Each account holds one key at a time; issuing a new key immediately invalidates the previous one. There are no OAuth flows or multi-token support.
- Bearer token in Authorization header
- Single active key per account
- Immediate invalidation upon key rotation
Latency and Speed
Generation speed averages roughly real-time, processing approximately 25 characters of text per second of audio output. Short phrases typically require a few seconds to synthesize, while longer texts scale linearly with duration.
Requests are processed sequentially, meaning high concurrency can introduce processing delays. The system does not offer streaming or partial audio delivery, requiring the full generation cycle before the file is returned.
Real-Time Performance
Synthesis operates at a consistent rate of approximately 25 characters per second for text length. This ratio holds for both tts-1 and tts-1-hd models, ensuring predictable duration calculations for smart home notifications and voice assistant responses.
Queue Wait Times
Since requests are processed one after another, concurrent usage may cause a request to wait in a queue. Under heavy load, generation time will exceed the standard real-time duration. This sequential processing ensures deterministic output but limits parallel throughput.
Network Dependency
Latency is primarily determined by the API endpoint at api.playbk3.com and the client's network connection. The API does not support WebSocket streaming, so the full audio file must be received before playback can begin. Network jitter affects total time-to-first-byte but not generation speed.
Data Privacy
HomeEcho prioritizes data integrity for IoT deployments by ensuring input text is processed solely for audio synthesis and not retained for model training. This guarantees that voice prompts sent to smart devices remain private and are not repurposed for improving the underlying engine.
Generated audio files are returned directly in the response and are not stored on our servers after delivery. Only request metadata, such as timestamp and character count, is retained strictly for billing purposes, keeping your voice data clean and audit-ready.
Input Text Usage
We use your input text exclusively to generate the requested audio file. The content is not ingested into our training datasets, ensuring your specific voice commands or notifications remain distinct from our model's learning corpus.
Audio Storage
Synthesized audio is streamed to your client and is not persisted in our storage systems. Once the response is delivered, the audio file is gone from our infrastructure, minimizing storage liabilities for high-volume IoT deployments.
Training Data
Input data is isolated from training pipelines. Unlike some competitors, we do not use your text inputs to refine voice models, ensuring deterministic output consistency for your smart home applications.
Cost Efficiency
HomeEcho charges $0.01 per 1,000 characters, with spaces and punctuation included in the count. Failed requests incur no charge, ensuring you pay only for successful synthesis. This flat rate simplifies budgeting for IoT applications without hidden tiered pricing structures.
Per-Character Costs
Each request processes up to 4,096 characters. For longer texts, split the input and join the resulting audio files. This predictable unit cost allows precise estimation of expenses for device notifications or voice assistant responses.
- $0.01 per 1,000 characters
- Spaces and punctuation count
- Failed requests are free
Trial Availability
Google sign-in accounts receive $0.50 in trial credit, valid for 7 days. This covers approximately 50,000 characters. Email-only accounts start with zero balance and require immediate top-up.
Minimum Spend
The minimum top-up is $5 via USDT on the TRON network. There are no monthly plans, seats, or minimum monthly spends. Balance does not expire.
Recommendation Summary
Selecting the right TTS API depends on your specific deployment constraints. HomeEcho offers deterministic voice generation suitable for standardized smart home notifications, while competitors may offer deeper ecosystem integration or advanced formatting options.
For Indie Developers
Start with the $0.50 Google trial. Use the OpenAI-compatible SDK by adjusting the base URL to api.playbk3.com/v1. This allows you to test voice presets like Aria or Felix without managing complex billing setups.
- Use the ready curl example on the dashboard.
- No minimum spend required.
- Simple USDT top-up for scaling.
For High Volume
Pre-split text into 4096-character segments to handle long documents efficiently. The per-character pricing at $0.01 per 1,000 characters remains consistent regardless of volume. Ensure your implementation handles the 60 requests per minute rate limit gracefully.
- Split input text programmatically.
- Join resulting audio files server-side.
- Monitor the /v1/voices endpoint for updates.
For Low Latency
Expect processing times similar to real-time speech for long texts, though short phrases may take a few seconds. Requests are queued sequentially, so high concurrency may introduce wait times. Optimize by reducing request frequency during peak loads.
- Requests are processed sequentially.
- Approx. 25 characters per second.
- No real-time streaming available.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This allows you to test voice presets like Aria or Felix without managing complex billing setups.
23 languages
Any voice speaks any of these languages.
