Frequently asked questions
What is the best voice for banking alerts?
Atlas offers a neutral, news-style tone ideal for clear financial updates. Sage provides a calm, instructional voice suitable for instructional banking steps. Both are available in all 23 supported languages. You can test each voice with a free trial credit to see which fits your app’s user experience best without committing to a specific style.
How long does integration take?
Integration is fast because the API follows the standard OpenAI speech endpoint shape. You only need to change the base URL to https://api.playa71.com/v1 and use your API key. Most developers can send their first audio file within minutes using the official OpenAI SDKs or simple HTTP requests, requiring minimal code changes.
Do I need a developer to integrate?
A developer is recommended to handle the API key management and text formatting. You will need to split long texts into 4096-character chunks and join the resulting audio files. While the endpoint is simple, managing rate limits (60 requests per minute) and error handling for production banking notifications typically requires technical implementation.
Can I use this with React Native?
Yes, you can use this API with React Native by making standard HTTP POST requests to the speech endpoint. Since the API uses standard JSON bodies and returns audio files, it works seamlessly with any platform that supports network requests. You do not need a specific mobile SDK, just your API key and the text input.
How do I handle errors in production?
Failed requests are free, so you can retry them without incurring extra costs. If synthesis fails, the charge is returned in full. You should implement retry logic for transient errors and ensure your app handles cases where the API returns an error status. Monitor your usage dashboard to stay within the 60 requests per minute limit.
What is the latency like?
Typical speed is roughly real time, generating about 25 characters of text per second. Short phrases may take a few seconds to process. Requests are handled sequentially, so under heavy load, a request might wait in a queue. For banking alerts, this latency is generally acceptable for asynchronous notifications rather than real-time conversation.
Can I send voice notifications via email?
The API returns an audio file, which you can attach to emails using your own email service. The API itself does not send emails directly. You generate the audio via the API, then use your existing notification infrastructure to deliver the audio file to your users’ inboxes, ensuring you control the delivery channel.
How do I test the API before launch?
Sign up with Google to receive $0.50 of free trial credit, valid for 7 days. This allows you to generate about 50,000 characters of audio. Use the provided curl examples on the /get-api-key/ page to test different voices, speeds, and formats. This lets you validate the audio quality and integration flow before going live.
Is the API secure?
Authentication is handled via a Bearer token in the Authorization header. Each account has one active key, which can be rotated instantly if compromised. Input text is used only for synthesis and not for training. While we do not offer specific enterprise certifications, the standard HTTPS connection ensures secure transmission of your banking data.
Can I customize the speed?
Yes, you can adjust the speed between 0.25 and 4.0 using the speed parameter in the request body. This allows you to make alerts faster or clearer depending on the context. For example, you might use a slower speed for complex financial instructions and a faster speed for simple balance updates.
What happens if the API is down?
We do not provide an SLA or uptime guarantee. If the API is unavailable, your requests will fail. Since failed requests are free, you can retry them later without cost. You should implement fallback mechanisms in your app, such as queuing notifications or using a different channel, to ensure users receive their alerts.
Do I need to store the audio files?
No, the API does not store generated audio files after returning them in the response. You are responsible for storing or discarding the audio files as needed. Input text is also not stored for training purposes. This reduces your storage costs and simplifies data management for your banking notification system.
Can I use multiple voices in one app?
Yes, you can use any of the 8 available voices in your application. You can assign different voices to different types of notifications, such as using Atlas for news-style alerts and Luna for storytelling features. The API supports all 23 languages, allowing you to tailor the voice and language to your user base.
How do I track costs?
Pricing is $10 per 1 million characters, or $0.01 per 1,000 characters. You can track your balance, usage, and characters remaining on the /get-api-key/ dashboard. Failed requests are free, so you only pay for successful generations. Top-ups via USDT are credited automatically, and your balance does not expire.
Why Use Voice Notifications in Fintech
Voice notifications transform static banking alerts into engaging, human-centric experiences. By delivering clear audio updates, fintech teams can significantly improve user engagement, ensuring critical messages are heard rather than just read in a crowded notification stream.
Improved user engagement
Audio alerts capture attention more effectively than text, especially for time-sensitive transaction updates. This human touch fosters a stronger connection between the user and the financial product, reducing churn in competitive markets.
Accessibility benefits
Voice notifications ensure your platform is inclusive for visually impaired or elderly users. Clear, natural-sounding speech removes barriers, allowing all users to manage their finances with equal ease and independence.
Compliance considerations
Audio records provide verifiable proof of notification delivery, supporting regulatory requirements for critical banking alerts. Ensuring clarity and accuracy in voice output helps maintain trust and meets standards for clear communication.
Step 1: Obtain API Credentials
Integrating VerbaClear’s TTS API for banking notifications begins with securing your API credentials. This initial step ensures your fintech product team can authenticate requests and begin delivering human-sounding voice alerts to users efficiently.
Once you have your credentials, you can proceed to configure your application to communicate with the API endpoint securely and reliably.
Create an account
Sign up on the /get-api-key/ page using your Google account or email and password. Each account is tied to the playa71.com storefront, so ensure you are registering on the correct site for your team’s needs.
Accounts created with email and password start without trial credit, while Google sign-ins receive $0.50 of credit valid for 7 days. This allows you to test the service immediately without immediate payment setup.
- Support Google or email/password sign-up
- $0.50 trial credit for Google sign-ins
- No monthly plans or minimum spend
Retrieve your key
After signing in, navigate to the dashboard to view your active API key. The interface displays your key, balance, and usage statistics.
You can issue a new key at any time, which immediately invalidates the old one. This feature simplifies key rotation for security purposes without disrupting your integration workflow.
- One active key per account
- Instant invalidation of old keys
- Dashboard shows balance and usage
Test connectivity
Use the provided curl example on the dashboard to verify your key works. Send a POST request to https://api.playa71.com/v1/audio/speech with a short test text to confirm you receive an audio file.
Ensure your request includes the correct Authorization header. Successful retrieval of the audio file confirms that your integration is ready for production use.
Note that new keys start with a limit of 10 requests per minute, increasing to 60 per minute after your first top-up.
- Use the dashboard curl example
- Verify Authorization header format
- Start with 10 requests/minute limit
Step 2: Choose the Right Voice
Choosing the right voice ensures your banking alerts sound appropriate for the context. Our eight presets are designed for clarity. Each voice is based on licensed recordings, providing a human touch without the cost of custom studio work.
Selecting the correct voice reduces cognitive load for users. A mismatched tone can make a simple balance check feel urgent or a security warning feel casual. We provide distinct personas to match specific notification types, from routine statements to critical security events.
Aria for narrations
Aria offers a warm, clear narrator tone ideal for monthly statements or detailed transaction summaries. Her clarity ensures that complex financial data is delivered without sounding robotic, making long-form voice notifications easier to digest for users reading on the go.
Use Aria when the message requires careful listening, such as explaining a new fee structure or summarizing weekly spending habits. Her steady pacing helps users retain information without feeling rushed by a high-energy delivery.
- Warm, clear narrator tone
- Ideal for monthly statements
- Best for detailed summaries
- Steady, readable pacing
Felix for promos
Felix brings an energetic, promo-ready voice to your marketing communications. His upbeat delivery is perfect for announcing new credit card offers, cashback events, or loyalty rewards. He cuts through notification fatigue with a positive, engaging tone that encourages user interaction.
Since Felix is designed for promos, he works well for time-sensitive offers where you want to convey excitement without sounding salesy. His voice aligns with modern fintech branding, keeping your promotional messages fresh and attention-grabbing.
- Energetic, promo-ready voice
- Perfect for new offers
- Conveys excitement effectively
- Ideal for time-sensitive rewards
Atlas for news
Atlas provides a neutral, news-style delivery suitable for broad market updates or system maintenance alerts. His tone is authoritative yet approachable, making him ideal for delivering factual information without adding emotional bias to the message.
Use Atlas for general notifications where clarity is paramount. Whether it’s a change in interest rates or a service update, Atlas ensures the message is delivered with professional consistency, reinforcing trust in your institution’s communications.
- Neutral, news-style delivery
- Authoritative yet approachable
- Ideal for market updates
- Ensures professional consistency
Step 3: Construct the Request
Constructing a valid request is straightforward, as the endpoint mirrors the OpenAI speech structure. Send a POST request to https://api.playa71.com/v1/audio/speech with your bearer token. This simplicity allows fintech teams to integrate quickly using standard HTTP libraries or the official OpenAI SDK by simply adjusting the base URL.
JSON body structure
The request body is a simple JSON object containing the text and configuration parameters. Ensure the input text does not exceed 4096 characters, as requests longer than this must be split by your application before sending.
Required fields
You must include model, input, and voice. The model accepts "tts-1" or "tts-1-hd" for identical engine performance. The input field holds your notification text, and voice selects one of the eight presets like Atlas for news or Aria for narrations.
- model: "tts-1" or "tts-1-hd"
- input: string up to 4096 characters
- voice: e.g., "atlas", "aria", "felix"
Optional parameters
Use response_format to choose between mp3, wav, opus, flac, or pcm. Adjust speed between 0.25 and 4.0 to fit your notification style. Setting language to an ISO code ensures correct pronunciation, especially for non-English Latin-script texts, preventing default English guessing errors.
- response_format: mp3, wav, opus, flac, pcm
- speed: 0.25 to 4.0
- language: ISO code (e.g., "en", "de")
Step 4: Handle the Response
Upon a successful POST request, the API returns the generated audio file directly in the response body. Depending on your chosen response_format, this will be an MP3, WAV, Opus, FLAC, or PCM file. Your banking application can immediately pipe this binary data to the user’s device or store it for later playback.
Audio file retrieval
Retrieve the audio payload from the HTTP response. Since the endpoint returns the file directly, you do not need to make a secondary call to fetch the audio. Ensure your client library handles binary data correctly without corrupting the file headers.
- Support for MP3, WAV, Opus, FLAC, and PCM formats
- Direct binary stream in the response body
- No secondary fetch endpoint required
Error handling
If synthesis fails, the request is free, and the charge is returned in full. Handle network timeouts and API errors gracefully, retrying with exponential backoff if necessary. Remember that requests are processed sequentially, so under load, a request may wait in a queue, increasing latency.
- Failed requests incur no cost
- Charge is returned in full on failure
- Retry logic recommended for queue delays
Status codes
Standard HTTP status codes indicate the outcome. A 200 OK means the audio is ready. Use 400 for invalid parameters (e.g., text exceeding 4096 characters) and 429 if you exceed the rate limit of 60 requests per minute per key.
Implement logic to parse these codes and alert your engineering team if error rates spike, ensuring your notification system remains reliable for critical banking alerts.
- 200: Success, audio file included
- 400: Bad Request (invalid input)
- 429: Rate Limit Exceeded (60 req/min)
Step 5: Manage Rate Limits
Rate limits define your throughput ceiling. New accounts start with a modest 10 requests per minute, which increases to 60 requests per minute once you add funds. This baseline dictates how you structure your notification pipelines, especially when sending bulk alerts to user bases.
Monitoring usage
Track your request count closely. The API returns usage data, allowing you to predict when you might hit the 60 requests/minute limit.
Burst handling
Use exponential backoff when you receive 429 errors. Since requests are processed sequentially, a sudden spike in transaction alerts can quickly exhaust your per-minute quota. Pause briefly and retry to avoid wasting characters on failed attempts.
Queue management
Implement a message queue in your backend. This smooths out traffic spikes, ensuring your TTS engine receives a steady stream of requests within the 60 rpm limit rather than choking on a burst of simultaneous balance updates.

Step 6: Optimize for Scale
Scaling your banking notifications requires efficient handling of high-volume bursts and large text payloads. By batching requests and caching frequent alerts, you can optimize throughput and reduce redundant processing costs.
Our API supports flexible response formats and speed adjustments, allowing you to balance audio quality with generation speed for critical user alerts.
Batch processing
Group multiple short notifications into a single logical workflow if your integration allows queuing. Since requests are processed sequentially, batching ensures predictable throughput without exceeding the 60 requests per minute limit per key.
- Keep requests under 4096 characters each
- Use a single active API key
- Monitor the 60 req/min limit
Caching strategies
Cache audio files for static messages like 'Welcome to VerbaClear' or standard security warnings. This avoids re-generating identical audio for every user login, saving both characters and time.
- Store static audio in CDN
- Cache based on voice + text + format
- Invalidate cache only when text changes
Long text handling
Split texts exceeding 4096 characters into smaller parts. Generate each segment individually and join the resulting audio files on the client side. This ensures no data loss and maintains audio continuity for longer transaction details.
- Split input at 4096 characters
- Join audio files client-side
- Handle potential gaps between segments
Step 7: Test with Real Data
Testing with real transaction data ensures your voice alerts sound natural and urgent without sounding robotic. Banking users expect clear, human-sounding notifications that convey importance without causing anxiety. Using actual transaction amounts and merchant names in your tests helps verify that the TTS engine handles mixed alphanumeric strings correctly.
Proper testing also confirms that latency remains acceptable during peak hours. Since requests are processed sequentially, validating performance with realistic data volumes ensures your system won’t bottleneck when sending high-volume alerts to thousands of users simultaneously.
Transaction alerts
Use diverse transaction types in your test data: purchases, refunds, and recurring payments. Verify that the voice clearly articulates numbers, currency symbols, and merchant names. For example, test a $14.99 coffee purchase to ensure the speech remains crisp and understandable.
Ensure the tone matches the transaction type. Use Felix for energetic promo alerts or Atlas for neutral transaction confirmations. This variety helps users distinguish between different alert types quickly.
- Test with various currency formats and decimal places
- Verify clarity of merchant names and transaction IDs
- Ensure tone matches transaction urgency
Balance notifications
Balance alerts often involve larger numbers and specific phrasing. Test voices like Sage or Luna for a calmer, more instructional tone when reporting account balances. This helps users process financial information without feeling rushed.
Check how the API handles edge cases like zero balances, negative overdrafts, or very large sums. Ensure the audio output remains clear and doesn’t distort on extreme values.
- Test with zero, negative, and large balance values
- Verify clarity of spoken numbers and percentages
- Ensure calm tone for informational updates
Security warnings
Security alerts require a sense of urgency without panic. Use Atlas or Orion for a neutral, authoritative tone that conveys importance. Test alerts for failed logins, unusual locations, or PIN changes to ensure the message is clear and actionable.
Verify that the audio file is generated quickly enough to be relevant. Since short phrases take a few seconds, ensure your system accounts for this latency when triggering real-time security alerts.
- Test urgency levels for different security threats
- Verify quick generation time for real-time alerts
- Ensure clear call-to-action in audio output
Step 8: Monitor Performance
Monitoring ensures your banking alerts remain timely and clear. Since requests are processed sequentially, latency can vary during high-volume periods. Track response times to identify when queues build up, ensuring critical transaction notifications aren't delayed beyond acceptable thresholds for your users.
Latency checks
Measure the time from request submission to audio file receipt. Short phrases typically take a few seconds, while longer texts generate at roughly real-time speed (25 characters per second). If your notification service experiences delays, check if your 60 requests-per-minute limit is being approached.
- Monitor average response time per request.
- Track queue wait times during peak hours.
- Set alerts for latency spikes exceeding 5 seconds.
Quality assessment
Periodically sample generated audio to verify clarity and tone. Ensure voices like Atlas sound neutral for balance updates or Felix remains energetic for promo alerts. Listen for consistent pronunciation across different languages and dialects supported by the API.
- Sample 1% of generated audio files weekly.
- Verify voice consistency across different message types.
- Check for audio artifacts or cuts.
Cost tracking
Track character consumption against your balance. Each request charges based on character count, including spaces. Failed requests are free, so monitor error rates to ensure you aren't paying for invalid inputs. Use the dashboard to view usage history and predict future costs.
- Monitor daily character usage.
- Track failed request rates.
- Compare cost per notification type.
Step 9: Deploy to Production
Moving your banking TTS integration from staging to production requires a careful approach to ensure reliability for high-volume notification systems. By using the OpenAI-compatible SDK with our base URL, you can deploy voice alerts that meet your specific latency and quality standards without introducing unnecessary complexity.
Staging environment
Verify your integration using the free trial credit ($0.50) or a small top-up of USDT on TRON. Test endpoints like /v1/audio/speech with various voices such as Atlas for news or Felix for promos to ensure the audio output matches your compliance requirements before scaling.
Rollout strategy
Adopt a phased rollout, starting with non-critical alerts like balance notifications. Monitor the 60 requests per minute limit per key, ensuring your system handles the queue effectively if load increases. Since requests are processed sequentially, plan for potential delays during peak usage.
User feedback
Collect feedback on audio clarity and speed (0.25–4.0 range) to refine voice selection. Support is available via email for any integration issues. Remember that generated audio is not stored, so focus on real-time user experience during this phase.
Maintain and Update
Keeping your integration healthy requires routine maintenance of credentials and monitoring support channels. Since each account holds only one active key, rotating it ensures security without downtime, as the previous key stops instantly upon issuing a new one.
Key rotation
Issue a new key via the dashboard at any time. The old key becomes invalid immediately, so update your notification service configuration promptly to prevent any interruption in alert delivery.
Voice updates
Voice presets like Aria, Felix, and Atlas remain constant. Check the /v1/voices endpoint regularly to confirm available options for your fintech alerts.
Support channels
Support is available by email only. There is no live chat or phone number. For urgent billing or integration issues, contact us directly via email with your account details for faster resolution.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Each voice maintains a consistent tonal character—such as upbeat, calm, or deep—throughout the generation.
23 languages
Any voice speaks any of these languages.
