Frequently asked questions
What is the difference between tts-1 and tts-1-hd?
Both models use the same engine and price. You can choose either 'tts-1' or 'tts-1-hd' in your request without any cost difference. They offer identical audio quality and processing speed, allowing you to switch between them freely based on your preference without affecting your billing or integration logic.
Can I use this API for live news broadcasts?
The API is not designed for live, real-time broadcasting. Requests are processed sequentially, so under load, a request may wait in a queue. While typical speed is roughly real-time for long texts, short phrases take a few seconds. It is better suited for pre-generated audio segments rather than live streaming feeds.
How accurate is the language detection?
If you do not set the 'language' field, 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 system supports 23 specific languages, including English, Spanish, and German.
Are the voices suitable for formal news reading?
Yes, specifically 'atlas' (neutral, news) and 'sage' (calm, instructional). Other voices like 'aria' or 'milo' offer different tones. All voices are based on licensed recordings from LibriTTS-R and LibriVox. They provide clear, broadcast-quality audio suitable for digital news and newsletters without requiring enterprise contracts or custom voice design.
Do I need a developer to integrate this API?
A developer is recommended but not strictly required. The API follows the standard OpenAI request shape, so official OpenAI SDKs work by changing the base URL. You can also use plain HTTP or ready curl examples provided on the dashboard. Basic programming knowledge helps manage rate limits and character splitting for long articles.
What happens if the API goes down?
There is no SLA or uptime guarantee. If synthesis fails, the charge is returned in full. Request metadata is kept for billing purposes. Since there is no enterprise contract, you should monitor your usage and balance. Support is available via email only, so plan for potential delays during outages without guaranteed recovery times.
Can I use my own text for the API?
Yes, you can send any text up to 4096 characters per request. Input text is used only to produce the audio and is not used for training. You can split longer texts into parts and join the audio files. Generated audio is returned in the response and not stored by the service.
Is there a limit on the number of voices?
There are exactly 8 voice presets available: aria, nova, luna, sage, felix, milo, atlas, and orion. You can choose any of these for each request. Each voice can speak any of the 23 supported languages. There is no limit on how many times you can switch between these voices in your API calls.
Do I need to pay for each language separately?
No, pricing is based on character count, not language. The rate is $10 per 1 million characters. All 23 supported languages, including English, Chinese, and Arabic, use the same voice engine and price. You do not need separate subscriptions or keys for different languages; just specify the language code in your request.
How long does it take to generate audio?
Typical speed is roughly real-time, about 25 characters of text per second of generation. Short phrases take a few seconds. Requests are processed one after another, so under load, a request may wait in a queue. Longer texts are split into parts, which may add slight delays between segments.
Can I listen to the audio before purchasing?
You can use the free trial credit of $0.50 (about 50,000 characters) valid for 7 days when signing in with Google. This allows you to generate and listen to audio samples. Accounts created with email and password start without trial credit, so you must top up via USDT on the TRON network to test the service.
Is the API compatible with WordPress?
There is no official WordPress plugin. However, you can integrate the API into WordPress using custom code or a plugin that supports HTTP requests. Since the API uses standard JSON bodies, you can write a simple PHP script to send text and receive audio files, storing them as media attachments in your site.
Do I need to credit the voices used?
The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. While the API does not enforce specific credit formatting, you may attribute the source voices as per the CC BY 4.0 license if required by your usage context. Each file also carries an inaudible synthetic watermark.
Can I use the API for political news?
Yes, the API does not restrict content type. You can generate audio for political news, opinions, or scripts. The voices like 'atlas' (neutral, news) are suitable for this. There are no emotion tags or custom voice cloning, so the tone remains consistent with the selected preset regardless of the political subject matter.
Voice quality for news
PressSynth offers a voice model optimized for consistent, professional audio output suitable for digital news delivery. The API provides eight distinct voice presets, each designed with specific tonal characteristics to match various broadcasting needs, from energetic promos to calm storytelling.
For news-specific applications, the Atlas voice is the primary recommendation due to its neutral, authoritative tone. This preset delivers clear, steady narration ideal for reading scripts or daily news roundups.
Atlas voice for neutral news
The Atlas voice is engineered for a neutral, news-read style. It avoids dramatic inflections, providing a steady and professional delivery that mimics standard broadcast journalism. This makes it suitable for delivering factual content.
Publishers can use Atlas for daily news scripts, breaking news alerts, or general informational content. Its consistent tone ensures that the focus remains on the information being conveyed, making it a reliable choice for routine audio generation tasks.
- Neutral, authoritative tone suitable for factual reporting
- Consistent delivery without dramatic emotional shifts
- Ideal for daily news scripts and breaking news alerts
- Works well for general informational content
Clarity and enunciation
The TTS engine prioritizes clear enunciation, ensuring that every word is distinct and easy to understand. This is crucial for news audio, where listeners need to catch details quickly. The voices handle standard sentence structures and proper nouns.
While the system does not offer phoneme-level customization, the default enunciation is sufficient for most news content. Users can adjust the speed (0.25–4.0) to ensure optimal clarity, especially for faster-paced breaking news updates.
- Clear, distinct word delivery
- Handles complex sentences and proper nouns well
- Adjustable speed (0.25–4.0) for optimal clarity
- No need for phoneme-level customization
Watermark identification
Every audio file generated by PressSynth includes an inaudible watermark that identifies it as synthetic speech. This helps maintain transparency with listeners, distinguishing automated content from human-recorded broadcasts. The watermark is embedded directly into the audio file.
This feature is particularly useful for news organizations that need to disclose the use of AI-generated audio in their publications. It ensures compliance with transparency standards while maintaining the professional quality of the broadcast.
- Inaudible watermark in every generated file
- Identifies audio as synthetic speech
- Maintains transparency with listeners
- Helps comply with disclosure standards
Language support for global news
PressSynth supports 23 languages, enabling publishers to reach diverse audiences without needing separate voice actors for each market. The API handles Latin-script text by defaulting to English, but you can explicitly set the language ISO code for other languages to ensure accurate pronunciation and intonation.
While voices are not region-specific, they can speak any supported language with consistent quality. This flexibility allows you to maintain a uniform audio brand across multiple global editions of your newsletters or digital publications.
23 languages available
The 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. Each voice preset can deliver speech in any of these languages.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Accents and dialects
Voices do not have fixed regional accents; they adapt to the language setting you provide. For non-English Latin-script text, always specify the language parameter to avoid defaulting to English pronunciation. This ensures that French, German, or Spanish content sounds natural and accurate.
Non-supported languages
Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, and other languages are not supported. If your audience requires these, you will need to integrate a different solution or use a workaround. Always verify language support before launching localized campaigns to avoid unexpected pronunciation errors.
Technical reliability
Reliable generation is essential for news cycles. PressSynth delivers consistent results, though infrastructure guarantees differ from enterprise contracts.
No SLA guarantees
We do not offer uptime Service Level Agreements or enterprise contracts. The API is designed for high availability, but specific uptime percentages are not guaranteed.
Queueing under load
Requests are processed sequentially. During high traffic, your request may wait in a queue before synthesis begins. This ensures fair resource distribution but adds latency during peak hours.
Real-time generation speed
Generation speed is approximately real-time, processing about 25 characters of text per second of audio. Short phrases may take a few seconds due to initialization overhead.
- Rate limit: 60 requests per minute per key
Integration with news workflows
Publishers can automate audio generation by integrating the PressSynth API directly into their content management systems or publishing pipelines. Using the standard OpenAI-compatible Python SDK, developers can send article text and receive audio files.
The API’s structure mirrors familiar industry standards, allowing existing Python scripts to be adapted quickly. This reduces the engineering overhead required to add audio capabilities to digital newsletters or news feeds without requiring deep specialization in audio processing.
API keys provide secure access to the generation engine. Each account maintains one active key, which can be revoked and replaced instantly if compromised, ensuring tight control over who can trigger audio synthesis within your newsroom’s infrastructure.
Python automation
Publishers can automate audio generation by integrating the PressSynth API directly into their content management systems or publishing pipelines. Using the standard OpenAI-compatible Python SDK, developers can send article text and receive audio files with minimal code, ensuring that script-to-audio workflows remain straightforward and reliable.
The API’s structure mirrors familiar industry standards, allowing existing Python scripts to be adapted quickly. This reduces the engineering overhead required to add audio capabilities to digital newsletters or news feeds without requiring deep specialization in audio processing.
- Use the official OpenAI Python SDK by updating the base_url
- Send JSON bodies with text, voice, and format parameters
- Automate batching for long articles by splitting text under 4096 characters
API key management
API keys provide secure access to the generation engine. Each account maintains one active key, which can be revoked and replaced instantly if compromised, ensuring tight control over who can trigger audio synthesis within your newsroom’s infrastructure.
Keys are issued upon account creation and can be regenerated at any time. When a new key is issued, the previous one becomes invalid immediately, preventing unauthorized usage from old scripts or distributed team members.
- One active key per account
- Instant invalidation of old keys
- Simple Bearer token authentication
Cost-effective scaling
The pay-per-character model eliminates the need for expensive enterprise contracts or monthly seat licenses. Publishers only pay for what they generate, making it easy to scale audio production for daily newsletters or breaking news alerts without predicting volume.
With a clear cost of $0.01 per 1,000 characters, budgets remain predictable. Failed requests are refunded, ensuring you do not pay for wasted generation cycles during high-volume publishing periods.
- No minimum spend or monthly commitments
- $0.01 per 1,000 characters
- Refunds for failed requests
Content usage rights
PressSynth ensures that your input text is used strictly for the purpose of generating audio files and is not utilized to train our models. This guarantees that your editorial content remains private and is not repurposed for other AI developments.
Every audio file produced carries an inaudible watermark, clearly marking the output as synthetic speech. This allows your audience to distinguish between human-recorded broadcasts and AI-generated audio, maintaining transparency in your news delivery.
Input text privacy
Text sent to the API is processed solely to create the audio file. We do not retain the input text for training purposes, ensuring your articles and scripts remain confidential and are not used to improve the engine’s voice models.
Synthetic speech watermark
Each generated audio file includes an inaudible watermark that identifies it as synthetic speech. This feature helps maintain trust with your listeners by allowing them to recognize that the voice is AI-generated, which is increasingly important in digital news environments.
Commercial usage allowed
You retain full rights to the audio files generated for your broadcasts. The output is free from additional licensing fees.
Comparison with traditional methods
Switching from human voice actors to an API changes the economics of audio production. Traditional studio sessions require booking time, hiring talent, and editing hours of raw audio. PressSynth generates speech in seconds, removing the logistical bottleneck of scheduling and post-production for routine news segments.
The cost difference is stark. While a professional voice actor might charge per hour or per project, PressSynth charges $0.01 per 1,000 characters. This allows publishers to produce unlimited daily updates without worrying about rising labor costs or minimum session fees for short clips.
Speed vs human voice actors
Human narration requires scheduling, recording, and editing, often taking hours for a single update. PressSynth generates audio quickly. For breaking news, this speed ensures audiences receive updates while they are still relevant, rather than waiting for a studio to clear.
Longer articles generate at roughly real-time speed, meaning a 10-minute article takes about 10 minutes to process. Requests queue under load, but for standard daily content, the turnaround is significantly faster than coordinating a voice actor's availability.
- [
Cost efficiency
Traditional voice-over work incurs fixed costs per session or hourly rates, which add up quickly for high-volume publishers. PressSynth’s pay-per-character model means you only pay for what you use. Failed requests are free, ensuring you never pay for errors.
At $10 per million characters, the cost is predictable and low. There are no monthly subscriptions or minimum spend requirements, making it ideal for variable news cycles where some days require extensive coverage and others are quiet.
- [
Scalability for daily news
Publishing daily news roundups or newsletter audio versions requires consistent, reliable output. The API handles volume seamlessly; you can generate hundreds of files in a day without negotiating additional contracts. This scalability supports growing audiences without linear cost increases.
The API operates continuously. You can schedule automated scripts to generate audio for early morning broadcasts or late-night updates without human intervention, ensuring consistent content delivery.
- [
Limitations for broadcast
PressSynth delivers reliable, clear audio for news workflows, but it lacks the expressive nuance of human broadcast talent. Understanding these constraints helps publishers plan their audio strategies effectively without overpromising on production values.
No emotion tags
The API does not support SSML or emotion tags. Voices like Atlas remain consistent in tone, which suits neutral news but lacks the dynamic range of a human anchor for dramatic stories.
No custom voice cloning
You cannot upload a recording to clone a specific journalist’s voice. The eight preset voices are fixed, meaning you cannot replicate a branded on-air personality without hiring them.
Standard audio formats only
Output is limited to mp3, wav, opus, flac, and pcm. There is no real-time streaming or WebSocket support, so you must generate full audio files before broadcasting them.

Use cases in news media
Automating audio production allows newsrooms to distribute content faster without hiring additional voice talent. By integrating PressSynth into existing CMS workflows, publishers can convert written articles into broadcast-ready audio files on demand.
Daily news roundups
Generate consistent audio summaries of daily articles using the Atlas voice for a neutral, professional tone. This ensures listeners receive uniform narration across multiple segments without variability between human narrators.
- Consistent narrator voice across all daily segments
- Automated conversion of CMS articles to audio
- Neutral delivery suitable for standard news reading
Breaking news alerts
Deliver immediate audio updates when stories develop. With API integration, text drafts can be converted to speech quickly, ensuring audiences receive timely information even during high-volume news cycles.
- Rapid conversion of breaking text to audio
- Immediate distribution via email or app notifications
- Scalable generation for multiple simultaneous alerts
Newsletter audio versions
Add an audio track to weekly newsletters to accommodate readers who prefer listening. Each article in the newsletter can be synthesized separately or combined, maintaining clear separation of topics.
- Enhanced engagement for audio-preferring subscribers
- Easy integration with existing email platforms
- Support for multiple languages for global audiences
Technical requirements
PressSynth operates entirely as a cloud-based API, requiring a stable internet connection for every request. Since synthesis happens on remote servers, your local infrastructure does not need specialized hardware, but latency depends on network stability and server queue times.
Authentication relies on a single Bearer token passed in the Authorization header. You can generate this key via your dashboard and rotate it instantly if compromised, ensuring your news pipeline remains secure without complex enterprise certificate management.
Internet connection needed
The API processes requests on remote servers, so a consistent internet connection is mandatory. Each request must reach the endpoint and receive the audio file before proceeding. Network interruptions will cause request failures, though failed attempts incur no charge.
Python SDK dependency
While you can use any HTTP client, the official OpenAI Python SDK is recommended for seamless integration. Simply change the base_url to https://api.98wi.info/v1 and use your PressSynth key. This allows you to leverage existing news automation scripts with minimal code changes.
Standard HTTP libraries like requests also work perfectly for simple scripts. The request shape matches OpenAI’s speech endpoint, making it easy to adapt existing news aggregation pipelines.
- Install via
pip install openai - Set
base_urltohttps://api.98wi.info/v1 - Pass key in
Authorization: Bearerheader
API key security
Each account holds one active key, which can be issued and rotated at any time from the dashboard. When a new key is generated, the old one becomes invalid immediately. This simplifies security management for news teams without needing complex key rotation schedules.
Keep your key in environment variables or secure configuration files within your automation scripts. Avoid hardcoding it in public repositories, as it grants full access to your character balance and generation capabilities.
- Issue new key instantly via dashboard
- Old key invalidates immediately upon rotation
- Use environment variables for scripts
Customer support for publishers
Publishers relying on PressSynth for daily news cycles need straightforward assistance when API integrations require tuning or billing questions arise. Our support structure is designed to minimize downtime without the overhead of enterprise-level phone trees, ensuring that technical issues are addressed efficiently by specialists who understand the API’s specific constraints and capabilities.
Email support channel
All inquiries are handled via email, allowing for detailed technical context and asynchronous communication that fits well with distributed newsroom teams. You can reach our support team for account issues, rate limit clarifications, or integration questions at any time, with responses provided as quickly as possible given the queue-based nature of our generation service.
- Direct email contact for technical and billing queries
- Asynchronous workflow suitable for global newsrooms
- Responses provided without wait times for live agents
Documentation resources
The /docs/ page provides comprehensive guides for integrating the TTS API into Python-based news automation pipelines. It includes detailed explanations of request fields, voice selection, and error handling, ensuring your development team can resolve most technical questions independently without needing to wait for a support response.
- Complete API reference for all endpoints
- Python integration examples and code snippets
- Guides for voice selection and language settings
No phone support
PressSynth does not offer a phone number or live chat, keeping operational costs low and passing those savings on to publishers. This model suits technical teams who prefer written records of support interactions and can manage issues through email and documentation rather than real-time voice calls.
- No dedicated phone line for customer calls
- No live chat widget on the site
- Support channel limited to email only
Pricing suitability
PressSynth charges $0.01 per 1,000 characters, making it highly predictable for newsrooms with variable daily output. Failed requests are free, ensuring you only pay for successful audio generation. This pay-per-character model aligns well with the fluctuating volume of news articles and newsletters.
There is no minimum spend or monthly subscription required. You can scale usage up or down without penalty, paying only for the characters you actually process. This flexibility supports both daily briefings and occasional special reports.
Pay-per-character model
The rate is $10 per 1 million characters, or $0.01 per 1,000 characters. Spaces and punctuation count toward this total. This straightforward pricing helps publishers forecast costs for regular audio content production without hidden fees.
- $0.01 per 1,000 characters
- $10 per 1 million characters
- Spaces and punctuation count
No minimum spend
No monthly plans or subscriptions are required. There is no minimum spend, allowing newsrooms to start small and scale as needed. Balance does not expire, so unused credits remain available for future projects.
Free trial availability
Signing in with Google provides a $0.50 free trial, valid for 7 days. This allows publishers to test the API with approximately 50,000 characters before committing to top-ups via USDT on the TRON network.
Accounts created with email and password do not include trial credit. Top-ups start at $5 minimum and are credited automatically within minutes.
- $0.50 trial credit
- 7-day validity
- Google sign-in required
Data privacy considerations
PressSynth ensures that the input text you submit is used solely for the purpose of generating the requested audio file and is not retained for model training or future development. This distinction guarantees that your proprietary news content remains distinct from the data used to improve the underlying engine.
Text not used for training
Unlike many consumer-grade TTS services, we do not repurpose your submitted articles, scripts, or newsletter content to train or refine our voice models. Once the audio is generated and returned in the response, the original text input is not stored for long-term model improvement purposes.
No audio storage
The generated audio files are delivered directly in the API response and are not stored on our servers after transmission. This ensures that your broadcast-ready audio clips remain ephemeral, reducing the risk of unauthorized access to your published news segments or exclusive reporting.
Metadata retention
We retain minimal request metadata, including the timestamp, character count, and status of each synthesis call. This data is kept strictly for billing accuracy and to verify that charges align with the characters processed, ensuring transparent accounting for your pay-per-character usage.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. The service provides eight preset voices based on licensed recordings from LibriTTS-R and LibriVox.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- Creative Commons Attribution 4.0
- LibriTTS-R corpus (OpenSLR 141)
- openai-python on GitHub
- IETF RFC 7845: Ogg Opus
- ISO 639 language codes
- OpenAI API reference: create speech
- LibriTTS-R paper (arXiv 2305.18802)
- openai-node on GitHub
- LibriVox public-domain audiobooks
- curl documentation
- FFmpeg documentation
