Frequently asked questions
Is the audio quality comparable to Google?
VoxCove uses licensed recordings from LibriTTS-R and LibriVox, delivering synthetic speech with inaudible watermarks. While we do not claim to be the best, our engine supports 23 languages and eight distinct voice presets. The audio is clear and suitable for professional radio spots, with speed adjustable from 0.25x to 4.0x for precise timing.
Can I use the same voices for multiple clients?
Yes. The API does not restrict voice usage per client. You can assign any of the eight presets—such as Aria or Felix—to multiple campaigns. Each generated file is a standard audio format (mp3, wav, etc.) that you can deliver directly to your clients without additional licensing constraints from the API side.
How do I handle multi-voice ads?
Send separate POST requests for each voice segment. Split your script into parts, assign the desired voice (e.g., Milo for one line, Atlas for another), and receive individual audio files. Your client-side logic or a simple script can join these files to create the final multi-voice radio spot.
Is the API suitable for live radio?
The API is designed for batch processing, not real-time streaming. Requests are processed sequentially, and generation takes roughly real-time duration (25 characters per second). For live radio, pre-generate your audio files during off-peak hours. Use the /v1/audio/speech endpoint to prepare your spots in advance.
What is the cost for a 30-second ad?
Pricing is $0.01 per 1,000 characters. A 30-second ad typically contains ~250 characters. This costs approximately $0.00125 per ad. If you generate 1,000 ads, the total cost is $1.25. Failed requests are free, so you only pay for successfully generated audio files.
Do I need a developer to integrate?
Integration requires sending JSON payloads via HTTP POST. While the API mimics OpenAI’s format, you need to handle base_url configuration and key management. If your agency has a developer familiar with basic API calls, integration is straightforward. Otherwise, you may need to hire technical support for initial setup.
Can I use the API for podcasts?
Yes. The API supports long texts up to 4096 characters per request. You can split longer scripts into parts and join the resulting audio files. The voices like Luna or Sage are well-suited for storytelling. Since audio is not stored, you manage your own podcast hosting infrastructure.
Are the voices licensed for commercial use?
The underlying recordings are from LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. The generated synthetic audio carries an inaudible watermark. You can use these files commercially for radio ads and podcasts without additional per-seat fees, as the API charges only by character volume.
How long does it take to generate audio?
Generation takes roughly real-time duration: about 25 characters of text per second. Short phrases take a few seconds. Under load, requests may wait in a queue since they are processed sequentially. For a 30-second ad, expect a few seconds of processing time plus queue latency.
Can I batch process hundreds of ads?
Yes. The API handles high-volume processing. You can split long texts into 4096-character chunks and process them sequentially. With a rate limit of 60 requests per minute, you can efficiently batch process hundreds of ads. Monitor your usage via the dashboard to ensure you stay within limits.
Is there a free trial available?
Yes. Signing in with Google provides $0.50 of trial credit, valid for 7 days. This covers approximately 50,000 characters. Accounts created with email and password do not receive trial credit. You can test the API’s quality and integration before committing to a top-up.
What payment methods are accepted?
We accept USDT on the TRON network (TRC-20) only. The minimum top-up is $5. Credits are added automatically within minutes of transfer. We do not accept credit cards, PayPal, or invoices. There are no monthly plans or subscriptions; you pay only for the characters you use.
Do I need to pay for each voice model?
No. The model field accepts either 'tts-1' or 'tts-1-hd', which use the same engine and price. All eight voice presets are included in the standard character rate. You do not pay extra for switching between voices like Felix or Orion within the same request or account.
Can I change my API key anytime?
Yes. You can issue a new API key from your dashboard at any time. The old key stops working immediately when the new one is activated. This allows you to revoke access if a key is compromised without affecting your balance or usage history.
Cost efficiency in advertising
Many TTS providers charge per second of audio generated, creating unpredictable costs for radio spots where delivery time varies based on voice speed and pauses. VoxCove bills per character of input text, decoupling cost from audio duration. This model ensures that a 10-second spot and a 15-second spot generated from similar text volumes incur comparable base costs, simplifying expense forecasting for production teams.
For short advertisements, the per-character model eliminates the penalty for extended pauses or slower, more natural speech rhythms. Agencies can adjust voice settings to enhance emotional impact without incurring extra fees based on extended audio length, ensuring that creative choices do not directly inflate production bills.
Per-character vs per-second billing
Many TTS providers charge per second of audio generated, creating unpredictable costs for radio spots where delivery time varies based on voice speed and pauses. VoxCove bills per character of input text, decoupling cost from audio duration. This model ensures that a 10-second spot and a 15-second spot generated from similar text volumes incur comparable base costs, simplifying expense forecasting for production teams.
For short advertisements, the per-character model eliminates the penalty for extended pauses or slower, more natural speech rhythms. Agencies can adjust voice settings to enhance emotional impact without incurring extra fees based on extended audio length, ensuring that creative choices do not directly inflate production bills.
- - Charge based on input text length, not audio duration - Consistent pricing regardless of voice speed settings - Clear calculation: $10 per 1 million characters
Reducing overhead for short spots
Short radio spots often require multiple iterations to perfect timing and tone. With per-character billing, agencies can experiment with different voice presets and speeds without worrying about accumulating high seconds-based costs. This flexibility encourages creative optimization without budget friction.
The straightforward character count also simplifies internal billing. Teams can estimate costs directly from the script word count, reducing the need for manual calculations or audits of generated audio files to verify duration-based charges.
- - No cost increase for slower, more expressive readings - Easy cost estimation from script length - Lower barrier for A/B testing voice options
Predictable monthly budgets
Agencies can forecast monthly TTS expenses with high accuracy. Since costs are tied to character volume, which is known before generation, budgeting becomes a simple multiplication of expected script length by the rate of $0.01 per 1,000 characters. This predictability aids in financial planning for large-scale campaigns.
Failed requests are free, removing the risk of paying for audio that does not meet quality standards. This ensures that only successful, usable audio files contribute to the monthly bill, further stabilizing cost projections for advertising departments.
- Fixed rate: $0.01 per 1,000 characters - No minimum spend or monthly subscriptions - Free charges for failed synthesis requests
Quality comparison
VoxCove’s engine delivers broadcast-ready clarity suitable for radio spots. Voices like Aria and Felix provide distinct tonal qualities, ensuring your message cuts through background noise without sounding synthetic.
Voice clarity and tone
The eight presets, ranging from the warm narration of Aria to the energetic delivery of Felix, allow precise matching to your brand’s sonic identity. Each voice is derived from licensed recordings, ensuring a natural baseline without the robotic artifacts common in older TTS systems.
Natural phrasing in ads
Generated audio handles standard ad copy with appropriate pacing. While the engine does not support SSML, the default phrasing is optimized for readability. Short phrases typically synthesize in a few seconds, allowing rapid iteration during creative testing.
Consistency across campaigns
Using the same voice preset across multiple spots ensures auditory continuity for your listeners. The API returns identical audio structures for identical inputs, maintaining uniform tone even when scripts vary slightly. This consistency builds recognition without requiring human voice actors for every variation.
Integration complexity
VoxCove eliminates integration friction by adhering to standard REST conventions. Agencies can ingest audio via simple JSON payloads without learning proprietary protocols. This reduces engineering time, allowing teams to focus on creative execution rather than infrastructure debugging.
The architecture mirrors industry standards, meaning existing devops pipelines require minimal adaptation. Our endpoints accept standard HTTP POST requests, ensuring compatibility with modern API gateways and legacy systems alike.
Standard REST API
Integration relies on straightforward HTTP POST requests to /v1/audio/speech. The JSON body accepts model, input, voice, and optional parameters like speed or language. This structure is familiar to most engineering teams, reducing the learning curve for new integrations.
Requests are processed sequentially, returning the audio file directly in the response. For longer scripts, the API handles character limits automatically, though splitting and joining segments remains the client's responsibility.
- HTTP POST to
/v1/audio/speech - JSON body with
model,input,voice - Synchronous audio return
OpenAI SDK compatibility
The API shape matches the OpenAI speech endpoint exactly. By simply updating the base_url to https://api.3phase.info/v1 and swapping the API key, existing OpenAI SDK implementations work immediately. This allows agencies to migrate scripts without rewriting core logic.
This compatibility extends to standard libraries used in Python, Node.js, and Go. Developers can retain their existing abstraction layers while switching the underlying voice engine, ensuring continuity in codebase maintenance.
- Same request shape as OpenAI
- Update
base_urlonly - Works with official OpenAI SDKs
No custom SDK required
VoxCove does not force a proprietary client library. Agencies can use any HTTP client capable of sending JSON, including curl, requests, or native fetch APIs. This flexibility prevents vendor lock-in and allows integration with diverse tech stacks.
For teams preferring direct control, the raw JSON structure is fully documented. This approach ensures that integration speed depends on your team's existing HTTP knowledge, not on mastering a new SDK.
- No proprietary client needed
- Works with any HTTP client
- Full JSON structure documented
Volume handling
Radio scripts often exceed the 4,096-character limit per request. The API handles this by splitting the input text into manageable segments, ensuring every character is processed correctly without manual intervention on your part.
Splitting long scripts
Send your full script up to the limit. The system divides longer texts into multiple requests, preserving the logical flow of your voiceover.
Joining audio segments
Each segment returns a separate audio file. You must concatenate these files to create the final spot. This ensures high fidelity for each part of the recording.
Managing request queues
Requests are processed sequentially. Under high load, a request may wait in a queue. Expect roughly real-time generation speeds, with short phrases taking a few seconds to complete.
- Limit: 60 requests per minute per key.
- Speed: ~25 characters per second on long texts.
Language support for global campaigns
VoxCove supports 23 languages, enabling agencies to produce localized radio spots without managing multiple voice providers. Each voice preset, from warm narrators to energetic promo voices, can deliver any of the supported languages with consistent quality.
This flexibility allows teams to maintain brand voice across markets while adapting to local linguistic nuances. The API handles language detection automatically, but explicit control ensures accuracy for campaigns requiring precise localization.
23 languages available
Our API supports 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, ensuring consistent audio quality regardless of the target market.
This broad coverage allows agencies to standardize their production pipeline. Whether producing a spot for Tokyo or São Paulo, the same voice presets deliver the same tonal characteristics, simplifying quality control across global campaigns.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Voice diversity
With eight distinct voice presets, agencies can match specific brand identities across different languages. Female voices like Aria and Luna offer warm, clear narration, while male voices like Felix and Atlas provide energetic or neutral tones suitable for news or promos.
This variety ensures that a brand’s audio identity remains recognizable whether the spot is in English, German, or Japanese. Each voice maintains its character traits, allowing for consistent brand recognition across diverse linguistic markets.
Automatic detection limits
If the language field is not set, the API guesses the language based on the script. Latin script text defaults to English, which may cause errors for non-English Latin-script text. For precise control, explicitly set the language parameter.
Supported languages are strictly limited to the 23 listed. Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, and others are not supported. Explicitly defining the language prevents misinterpretation and ensures accurate phonetic generation for international spots.
Technical requirements
The endpoint accepts standard HTTP POST requests to deliver audio synthesis jobs. You send a JSON payload containing your script and configuration parameters, and the API returns the resulting audio file. This method integrates seamlessly with existing ad tech stacks and automation tools without requiring proprietary clients.
Each request is processed sequentially, ensuring consistent billing and predictable delivery times for your radio spots.
HTTP POST requests
Send requests to https://api.3phase.info/v1/audio/speech. The endpoint handles the synthesis job and returns the audio data directly in the response body.
Requests are processed one after another. Under load, you may experience queuing delays, but the system maintains order and reliability for high-volume campaigns.
JSON body structure
The body must be valid JSON. Required fields include model ("tts-1" or "tts-1-hd"), input (your script text, max 4096 characters), and voice (one of the 8 presets).
Optional fields allow you to specify response_format (mp3, wav, opus, flac, pcm), speed (0.25–4.0), and language (ISO code). If omitted, the language is guessed, though setting it explicitly is recommended for non-English Latin scripts.
Authorization headers
Include your API key in the header: Authorization: Bearer tts-....
Each account holds one active key. You can issue a new key at any time; the old one stops immediately. Ensure your key is securely stored in your ad production environment.
Pricing transparency
Pricing is straightforward: $10 per 1 million characters, or $0.01 per 1,000 characters. Spaces and punctuation count toward this total. There are no monthly plans, seat fees, or hidden surcharges, ensuring your advertising budget remains predictable regardless of campaign volume.
No hidden fees
Unlike models that charge per API call or per word, we bill strictly on character count. This clarity prevents unexpected costs during high-volume production runs for radio spots.
Free failed requests
If synthesis fails for any reason, the charge is returned in full. You only pay for successfully generated audio files, protecting your ad spend from technical glitches.
No minimum spend
There is no minimum spend requirement. You can top up with a minimum of $5 via USDT (TRC-20) and use exactly what you need. Balances do not expire.

Voice selection for branding
Selecting the right voice is critical for radio spot branding. Our eight presets cover distinct tonal ranges, ensuring you can match the specific mood of your campaign without negotiating custom contracts.
The API offers four female and four male voices, each engineered for clarity and consistency. This diversity allows agencies to A/B test different brand personas efficiently.
8 distinct presets
We provide eight distinct voice presets, ranging from warm narrators to energetic promo styles. Each voice is optimized for commercial delivery, ensuring clear articulation even at high speeds.
- Aria: Warm, clear narrator
- Nova: Bright, upbeat
- Luna: Soft, storytelling
- Sage: Calm, instructional
- Felix: Energetic, promo
- Milo: Friendly, conversational
- Atlas: Neutral, news
- Orion: Deep, documentary
Male and female options
The selection includes four female and four male voices. This balance allows for versatile casting, from deep documentary styles to friendly conversational tones, without needing multiple voice actors.
Matching brand voice
Use these presets to maintain consistent brand identity across multiple radio spots. Since voices are based on licensed recordings, they offer a professional, grounded sound suitable for advertising.
Data privacy considerations
For advertising agencies, input text is used solely to produce the audio and is never used for training our models. Generated audio is returned in the response and not stored on our servers, ensuring your scripts remain confidential during processing.
We retain request metadata—time, character count, and status—exclusively for accurate billing and usage tracking. This minimal data retention supports transparent invoicing without storing the creative assets themselves.
Text not used for training
Input text is used only to produce the audio and is not used for training. This ensures your unique ad copy remains proprietary and is not absorbed into our model's training data.
No audio storage
Generated audio is returned in the response and not stored. Once the audio file is delivered to your application, we do not keep a copy, reducing liability and storage overhead.
Metadata retention
Request metadata (time, characters, status) is kept for billing. This allows us to verify usage and charge accurately without retaining the actual audio content or source text.
Migration steps
Migration begins by securing your unique API key on the /get-api-key/ page. This key authenticates your requests and tracks usage against your balance.
Next, update your SDK or HTTP client configuration. Point the base URL to https://api.3phase.info/v1 to route calls through our infrastructure.
Generate API key
Visit the /get-api-key/ page to create an account via Google or email. The dashboard displays your active key, current balance, and a ready-to-use curl example for immediate testing.
Update base URL
Change your OpenAI SDK base_url to https://api.3phase.info/v1. The request shape remains identical to the standard speech endpoint, requiring only the model name (tts-1 or tts-1-hd) and your new Bearer token in the Authorization header.
Test with sample scripts
Send a POST request with a short radio spot script to verify voice output and latency. Check the response for the audio file in your preferred format (mp3, wav, or opus) before integrating into your production pipeline.
Limitations to note
The API does not support SSML tags for precise pronunciation or emphasis control, so prosody is handled automatically by the model. This simplifies integration but reduces fine-grained editorial control over specific words or pauses within a single request.
Voice cloning is not offered; you cannot upload a reference audio to create a custom brand voice. You must choose from the eight available presets, which are based on licensed recordings. Each generated file includes an inaudible synthetic watermark.
No SSML support
The engine processes plain text only. You cannot use SSML tags like <break> or <emphasis> to adjust timing or tone for specific ad copy elements. The model applies natural phrasing based on the input text structure.
No voice cloning
Custom voice creation is not available. You cannot upload audio samples to generate a unique voice for your brand. You must select from the eight standard presets, which vary by tone and gender.
Rate limits apply
Each API key is limited to 60 requests per minute. New accounts start with a limit of 10 requests per minute until the first top-up is processed. Long scripts are split internally, but each segment counts as a request.
Under high load, requests may queue. Short phrases take a few seconds to generate, while longer texts process at roughly real-time speed. Plan your batch jobs accordingly to avoid throttling.
- 60 requests/minute per key after top-up
- 10 requests/minute for new accounts
- Requests queue during high load
Support and documentation
Documentation is hosted at https://3phase.info/docs/ and covers every endpoint, from voice lists to speech synthesis. We provide clear specifications for the JSON body structure, required headers, and response formats so your engineering team can integrate without guesswork.
Support is handled via email only. There is no live chat or phone line. We respond to technical queries directly, ensuring you get precise answers without waiting in a queue for general inquiries.
Email support only
All support tickets are managed through email. We do not offer phone support or live chat. This ensures that your technical questions are routed directly to engineers who understand the API architecture, reducing resolution time for integration issues.
Detailed API docs
The documentation at https://3phase.info/docs/ provides complete reference material. It details the request shapes for POST /v1/audio/speech and GET endpoints. You will find clear examples for handling rate limits, character splitting, and error codes.
We also list all 23 supported languages and 8 voice presets with their specific characteristics. This allows you to select the right voice for your radio spots without trial-and-error.
- Endpoint reference for /v1/audio/speech
- List of 8 voice presets
- Supported languages and formats
- Rate limit guidelines
Example code snippets
We provide ready-to-use code snippets for the most common workflows. These examples demonstrate how to send a POST request with the correct JSON body and authorization header.
You can adapt these snippets for your preferred language or framework. The examples cover basic speech generation as well as handling the response to retrieve the audio file.
The snippets assume you have already generated your API key from the /get-api-key/ page. They use standard HTTP libraries, so no proprietary SDK is required.
- cURL examples
- Python requests snippet
- JavaScript fetch example
- Error handling patterns
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Each voice is derived from licensed recordings, providing a consistent baseline for quality.
23 languages
Any voice speaks any of these languages.
