Frequently asked questions
Which voice is best for news articles?
Atlas is designed as a neutral, news-oriented voice suitable for standard reporting. For conversational segments, Milo offers a friendly tone. All eight voices support all 23 languages, allowing you to select the most appropriate tone for your specific editorial style and audience expectations.
How do I handle articles longer than 4096 characters?
Split the input text into parts of up to 4096 characters each. Send separate POST requests to the /v1/audio/speech endpoint for each segment. Receive the individual audio files and join them sequentially. This approach ensures compliance with the character limit per request while maintaining the integrity of the full article audio asset.
What is the best audio format for web players?
MP3 is suitable for web players due to broad compatibility and efficient compression. WAV, Opus, FLAC, and PCM are also available. MP3 is recommended for general playback, while WAV suits high-fidelity needs. All formats are supported by the API, allowing you to select the option that best fits your specific distribution requirements.
How do I track API usage costs?
Monitor your balance and character usage directly on the /get-api-key/ dashboard. The API charges $0.01 per 1,000 characters. Failed requests are free. Request metadata, including time, characters, and status, is retained for billing purposes. You can review your usage history and current balance at any time via the account page.
Can I use the API for podcast production?
Yes, the API supports all eight voices suitable for podcast narration. You can generate audio files in formats like MP3 or WAV. Since there is no SSML support, you must manage pacing and emphasis through text structure or post-production. The API is reliable for pre-recorded content but not designed for live, real-time streaming.
How does language detection work?
If the language field is omitted, the system guesses the language from the script. Latin script text defaults to English. For non-English Latin-script text, you should explicitly set the language parameter to ensure accuracy. The API supports 23 specific languages; any language outside this list will not be processed correctly.
What is the difference between MP3 and WAV?
MP3 is a compressed format, resulting in smaller file sizes suitable for web delivery. WAV is uncompressed, preserving higher audio quality but increasing storage and bandwidth requirements. Both are supported by the API. Choose MP3 for efficient distribution and WAV if you require lossless audio for professional editing or archival purposes.
How long does it take to generate audio?
Generation is roughly real-time, processing about 25 characters of text per second of audio. Short phrases may take a few seconds. Under load, requests are processed sequentially and may wait in a queue. Longer texts require splitting, so total time includes processing each segment individually before joining the final audio file.
Do I need a developer to integrate the API?
A developer is recommended to handle the JSON request structure, authentication headers, and audio file management. The API uses a standard OpenAI-compatible endpoint shape. You need to implement logic for splitting long texts and joining audio files. While the integration is straightforward, managing rate limits and error handling requires technical implementation.
Can I change the voice after generating audio?
No, the voice is determined at the time of the API request. Once the audio file is generated, the voice is fixed. To use a different voice, you must send a new request with the desired voice parameter. There is no post-generation voice modification or editing capability within the API itself.
Is the API suitable for live broadcasts?
No, the API is not suitable for live broadcasts. Requests are processed sequentially and may queue under load, introducing latency. There is no streaming or WebSocket support for real-time partial audio. It is designed for pre-recorded audio assets where latency is not critical. Use it for on-demand content rather than live feeds.
How do I handle failed requests?
If synthesis fails, the charge is returned in full. You do not lose credit for failed requests. You should implement retry logic in your application to resend the request after a brief delay. Monitor the response status code to determine if a request failed and trigger the appropriate error handling or retry mechanism.
What happens if I exceed the rate limit?
The rate limit is 60 requests per minute per key. If you exceed this, requests may be queued or rejected. New accounts start with 10 requests per minute. Exceeding the limit does not incur extra charges but may delay processing. Monitor your usage to stay within limits.
Can I use the API for multiple languages?
Yes, the API supports 23 languages including English, Spanish, German, and Japanese. Any voice can speak any of these languages. You can specify the language parameter in your request or allow the system to guess it. Ensure you use the correct ISO code for accurate pronunciation and language handling in your integrations.
Planning the Integration
Before integrating NewsAudio Reader, product teams must quantify the daily volume of articles and transcripts requiring conversion. Accurate character counting, including spaces and punctuation, ensures correct budgeting at $0.01 per 1,000 characters. This metric drives infrastructure decisions and prevents unexpected rate-limit delays during peak news cycles.
Defining specific use cases, such as accessibility compliance or multi-channel distribution, determines whether audio assets are generated on-demand or pre-processed. Teams should evaluate latency requirements, noting that short phrases take a few seconds while long texts generate at roughly real-time speed.
Assessing content volume needs
Calculate expected daily character volume to estimate API costs and select appropriate rate-limit strategies. Since requests process sequentially under load, high-volume batches may queue, affecting immediate availability.
Monitor character counts precisely, as every space and punctuation mark counts toward the $10 per million characters pricing model. Failed requests incur no charge, providing a safety net for testing.
- Count spaces and punctuation as characters
- Estimate daily generation volume
- Plan for sequential processing delays
Defining audio use cases
Identify whether audio serves accessibility needs (WCAG) or distribution across platforms. On-demand generation suits variable news cycles, while pre-processing ensures consistent delivery for scheduled broadcasts.
Consider latency constraints. Short phrases require a few seconds of generation time, whereas long articles approximate real-time speed. Plan server-side splitting if text exceeds 4,096 characters per request.
- Accessibility compliance (WCAG)
- Multi-channel distribution
- On-demand vs. pre-processed
Selecting voice presets
Choose from eight licensed presets based on content tone. For news, Atlas offers a neutral, professional delivery suitable for hard news updates. Milo provides a friendly, conversational tone for features or interviews.
Ensure voice selection aligns with the target audience's language. Any voice can speak any of the 23 supported languages, but setting the language parameter is recommended for non-English Latin-script text to avoid incorrect detection.
- Atlas: neutral, news
- Milo: friendly, conversational
- Set
languageparameter explicitly
Account Setup and Key Management
Newsroom teams begin by selecting a registration method on the /get-api-key/ page. Google sign-in provides immediate trial credit, while email accounts start without trial credit and require manual top-ups via USDT on the TRON network (TRC-20).
Google vs Email registration
Google authentication activates a $0.50 trial valid for seven days. Email-based accounts start without trial credit, requiring an immediate minimum $5 top-up to begin usage.
API key retrieval steps
The dashboard displays the active Bearer key immediately after login. Reissuing a key invalidates the previous one instantly, ensuring only one active credential exists per account.
- Navigate to the API key section on the dashboard.
- Copy the
tts-...string for your Authorization header. - Generate a new key if the old one is compromised.
Balance monitoring setup
The interface shows current balance, characters remaining, and usage history. Since billing is per character, monitoring ensures cost predictability for high-volume news transcripts.
- View real-time balance and character counts.
- Check usage logs for specific request timestamps.
- Monitor top-up status after TRC-20 transfers.

Technical Implementation
Configure the API base URL to https://api.playmomoo.com/v1 within your client environment. This endpoint handles all speech synthesis requests. Ensure your authentication header includes the Bearer token generated from your account dashboard.
Configuring base URL
Point your HTTP client to https://api.playmomoo.com/v1. This base URL ensures compatibility with standard OpenAI SDK structures. Verify that the domain resolves correctly before initiating authentication.
Setting up SDK client
Initialize the OpenAI SDK with the custom base URL. Pass your API key via the Authorization: Bearer header. The client automatically handles JSON serialization for the request body containing model, input, and voice parameters.
- Use standard HTTP libraries if preferred over SDKs
- Supports
tts-1andtts-1-hdmodels
Handling response objects
The API returns an audio file directly in the response body. Parse the binary data to store the audio asset. Check the HTTP status code to confirm successful synthesis before processing the audio stream.
- Status 200 indicates successful generation
- Response format matches requested
response_format
Voice Selection Strategy
Selecting the appropriate voice preset ensures audio assets align with editorial standards. Newsroom teams must distinguish between hard news delivery and feature storytelling to maintain audience trust and clarity.
The API provides eight distinct presets. Choosing correctly impacts listener retention and accessibility compliance without requiring custom voice design or cloning.
Neutral voice for news
Atlas offers a neutral, professional tone ideal for breaking news and factual reporting. Its consistent cadence ensures clarity during high-volume broadcast pipelines.
Atlas works across all 23 supported languages, making it suitable for international wire services distributing content globally without language-specific voice switching.
- Best for: breaking news, press releases, factual reporting
Conversational tone options
Milo provides a friendly, conversational style suitable for interviews, podcasts, or human-interest stories. It adds warmth to standard article conversions.
Sage serves as a calm, instructional alternative for explanatory content or guided tours of complex topics, ensuring listeners remain engaged.
- Best for: interviews, human-interest stories, explanatory content
Matching voice to content type
Match Aria for warm narratives, Luna for storytelling, and Felix for energetic promos. Orion suits deep, documentary-style features.
All voices support the same 23 languages. Use the language parameter to override automatic detection if the content script differs from the voice's primary language.
- Aria: warm narratives
- Luna: storytelling
- Felix: energetic promos
- Orion: documentaries
Format Optimization
Selecting the correct audio format balances fidelity against bandwidth constraints for news distribution. The API supports MP3, WAV, Opus, FLAC, and PCM via the response_format parameter. Each format serves distinct technical requirements for web players, archival storage, or efficient streaming.
MP3 for web delivery
MP3 is the standard for web delivery due to its universal compatibility and reasonable file size. It is suitable for most news articles where audio plays directly in a browser player.
WAV for high quality needs
WAV provides uncompressed, high-quality audio ideal for archival purposes or professional editing workflows. It ensures no loss in fidelity but results in larger file sizes compared to MP3.
Opus for bandwidth efficiency
Opus offers superior compression for the same perceptual quality, reducing bandwidth usage significantly. It is optimal for mobile users or regions with limited network infrastructure.
Handling Long Texts
News articles often exceed the single-request limit, requiring your backend to split input text into chunks of up to 4096 characters. Each chunk generates a separate audio file via the API. This approach ensures that no single request blocks the queue for too long during high traffic periods.
Splitting articles into parts
Divide the source text at natural boundaries, such as paragraph breaks, while ensuring no chunk exceeds 4096 characters. This prevents truncation and maintains logical audio segmentation. Each segment is submitted as an independent request to the speech endpoint.
Joining audio segments
The API does not join these segments; your backend logic must retrieve the individual audio files and concatenate them sequentially. Once all parts are downloaded, your system combines them into a single audio asset for distribution.
Managing character limits
Track character counts precisely, as spaces and punctuation count toward the 4096 limit. Failed requests are free, but successful ones consume your balance. Efficient splitting reduces queue wait times, as shorter requests process faster under load.
Error Handling and Retries
News platforms must handle transient API failures gracefully to ensure uninterrupted audio delivery. The API returns standard HTTP status codes, allowing your integration to distinguish between client errors (4xx) and server-side synthesis failures (5xx). Failed requests incur no charge, so verifying the response status is critical before logging costs or updating usage metrics.
Implementing robust retry logic ensures high availability without overloading the service. You should configure exponential backoff for 5xx errors, but avoid retrying 4xx errors like invalid input or missing keys. For rate limit errors (429), pause briefly before retrying to respect the 60 requests per minute limit per key (or 10 per minute before the first top-up).
Identifying failed requests
Check the HTTP status code in the response header. A 200 OK indicates successful synthesis. Any 4xx or 5xx code signifies a failure. Since failed requests are free, you can safely retry server errors without worrying about double-charging your balance.
Parse the JSON response body for detailed error messages. The API provides specific reasons for failures, such as invalid voice IDs or text exceeding the 4096 character limit. Use these messages to correct request parameters before attempting a retry.
- 200 OK: Audio generated successfully.
- 400 Bad Request: Invalid parameters or input.
- 401 Unauthorized: Invalid or expired API key.
- 429 Too Many Requests: Rate limit exceeded.
- 500-503 Server Errors: Temporary service issues.
Implementing retry logic
Use exponential backoff for transient 5xx errors. Start with a short delay and increase it with each retry attempt. This prevents overwhelming the server during peak loads. Limit the number of retries to three to avoid excessive latency in your audio pipeline.
Do not retry 4xx errors unless the issue is resolvable, such as a temporary network glitch. For invalid inputs, fix the data and retry immediately. Ensure your retry logic respects the per-key rate limit to avoid triggering 429 errors.
Managing rate limits
The API enforces a limit of 60 requests per minute per API key. Before the first top-up, this limit is stricter at 10 requests per minute. Monitor your usage metrics to stay within these bounds and avoid service interruptions.
Implement rate limit headers in your HTTP responses if available. If you receive a 429 error, wait for the duration specified in the Retry-After header before retrying. Use separate keys for different services if you need higher aggregate throughput.
- Standard limit: 60 requests per minute.
- Pre-top-up limit: 10 requests per minute.
- Retry-After header: Use for 429 errors.

Cost Management
Understanding character-based billing is essential for newsroom budgeting. Each character, including spaces and punctuation, counts toward the total. The standard rate is $0.01 per 1,000 characters, or $10 per million characters. Failed requests do not incur charges, ensuring you only pay for successfully generated audio assets.
To optimize costs, review your input text before sending it to the API. Removing unnecessary whitespace or redundant phrasing can reduce the character count without affecting audio quality. Since the API processes requests sequentially, efficient batching helps manage throughput during peak news cycles.
Tracking character usage
Monitor your dashboard to see real-time character consumption and remaining balance. The API returns usage metadata with each response, allowing your systems to log exact costs per article. This transparency helps finance teams forecast monthly expenses based on article volume and average word count.
Note that the free trial provides $0.50 of credit, valid for seven days. After this period, or if you registered via email without a trial, you must top up to continue generating audio. Regularly check your account balance to avoid service interruptions during breaking news coverage.
- $0.01 per 1,000 characters
- Failed requests are free
- Balance does not expire
Optimizing text length
Split long transcripts into segments under 4,096 characters to ensure reliable processing. The API joins these segments seamlessly, but keeping chunks manageable reduces queue wait times during high load. This approach also allows for more granular cost tracking per section of a report.
Use the speed parameter to adjust generation time, which indirectly affects server load but not character costs. Faster speeds do not change the price, but they may reduce the time your infrastructure holds the connection open.
Monitoring top-up cycles
Top-ups are processed via USDT on the TRON network (TRC-20). Transfers of at least $5 are credited automatically within minutes. There are no monthly subscriptions or minimum spend requirements, allowing you to fund the account only when needed.
Since there is no invoice system, keep records of your blockchain transaction hashes for accounting purposes. Balance remains available indefinitely, so you can top up during low-traffic periods to ensure availability during major news events without worrying about expiration.
Rate Limits and Costs
The rate limit is 60 requests per minute per key. If you exceed this, requests may queue, but costs remain tied to successful synthesis. Ensure your retry logic respects these limits to avoid unnecessary processing delays.
Use the GET /v1/voices endpoint to check pricing and format options before bulk generation. This helps you select the most cost-effective format, such as MP3, for web delivery, or WAV for archival purposes, without incurring unexpected fees.
Testing and Validation
Validate your integration by executing a basic request using the provided curl example. This confirms connectivity and key validity before processing full articles. Ensure the response returns the audio binary or error code correctly.
After receiving the file, verify audio quality and format consistency. Play the output to ensure clarity and check that metadata aligns with your newsroom’s technical standards for accessibility and multi-channel distribution.
Running curl examples
Use the provided curl command to send a POST request to the API base URL. Replace the placeholder key with your active token. Verify the response contains the audio binary or a JSON error object.
- Replace
tts-...with your actual API key - Set
modeltotts-1ortts-1-hd - Verify response headers for content-type
Validating audio output
Download the generated file and play it using a standard media player. Check for audio clarity, correct speed, and absence of artifacts. Confirm the file format matches your response_format request (mp3, wav, opus, flac, or pcm).
- Verify audio duration matches text length
- Check for consistent volume levels
- Confirm no clipping or distortion
Checking language detection
Test language auto-detection by sending Latin-script text without a language field. Verify the output defaults to English. For non-English Latin text, explicitly set the language parameter to ensure correct pronunciation and accent.
Accessibility Compliance
NewsAudio Reader supports WCAG 2.1 compliance by providing clear, machine-readable audio versions of text content. This allows news platforms to offer alternative formats for users with visual impairments or reading difficulties, ensuring broader accessibility across digital channels.
WCAG guidelines alignment
The API generates standard audio files that integrate seamlessly with existing web players. By converting article text into speech, newsrooms can meet accessibility requirements for digital publications without relying on third-party accessibility widgets.
- Supports WCAG 2.1 standards for audio alternatives
- Generates standard MP3 or WAV files compatible with screen readers
- Ensures consistent audio quality for reliable comprehension
Audio player integration
Delivered audio files can be embedded directly into article pages using standard HTML5 audio tags. This allows news platforms to maintain their existing design while adding an accessible audio layer to each story.
- Compatible with standard HTML5 audio elements
- Supports common web audio players and custom UIs
- Works with existing content management systems
User control features
Users can control playback speed, pause, and skip segments, enhancing the listening experience. News platforms can expose these controls to allow readers to adjust the audio to their personal preferences and reading pace.
- Adjustable playback speed (0.25x to 4.0x)
- Standard pause, play, and skip controls
- No additional SDK required for basic controls
Deployment Considerations
Deploying the NewsAudio Reader API requires accounting for sequential processing latency. Under load, requests queue, meaning long articles may take several seconds to synthesize. Plan your infrastructure to handle this synchronous delay without blocking user interfaces, ensuring that background workers manage the 4096-character splitting and joining logic efficiently.
Server load implications
Since requests process one after another during high volume, your backend must manage concurrency carefully. Avoid simultaneous bursts that exceed the 60 requests per minute limit per key, as this will increase synthesis time for longer texts.
Caching strategies
Cache generated audio files aggressively. Since the API does not store generated audio, you must persist the response on your own servers or object storage. Implement a cache key based on the text, voice, format, and speed to avoid redundant synthesis for identical content.
Content delivery networks
Distribute cached audio assets via a CDN to reduce latency for global news audiences. This offloads bandwidth from your origin server and ensures consistent playback quality across different regions, complementing the API’s real-time generation speed.

Maintenance and Updates
Once deployed, the NewsAudio Reader API requires minimal but consistent oversight. Since the service lacks an SLA or uptime guarantee, product teams should establish basic monitoring for request success rates and latency trends. Regularly reviewing usage dashboards helps anticipate scaling needs before volume spikes impact performance.
Periodic reviews ensure the integration remains aligned with evolving content strategies. While the core engine remains stable, checking for new voice presets or format updates can optimize output quality without requiring significant code changes.
Monitoring API status
Track request volumes and failure rates via the dashboard. Since support is email-only, maintain a log of recurring issues to streamline communication during peak news cycles.
- Set alerts for sudden rate limit breaches
- Log error codes for email support tickets
Updating SDK versions
The API uses standard OpenAI-compatible endpoints. Update your SDK client periodically to ensure compatibility with any minor field changes in the request schema. No custom SDK is required, so standard library updates suffice.
Reviewing voice options
Re-evaluate voice selection quarterly. As news formats shift, consider if neutral voices like Atlas remain optimal for hard news or if conversational tones like Milo better suit podcast-style segments.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. News outlets will amplify your launch, turning product updates into public conversation for your entire audience.
23 languages
Any voice speaks any of these languages.
