French Text to Speech API: Voice Selection & Integration Guide

By VoiceScope editorial team · Updated 2026-11-18

This guide explains how to integrate VoiceScope’s French text-to-speech API into iOS and Android apps. Freelance mobile developers will find concrete examples for using the OpenAI-compatible SDK, managing the 8 licensed voices, and handling character-based billing without infrastructure overhead.

You will learn the exact JSON request structure, supported audio formats, and performance metrics specific to French and multilingual apps. The page covers practical workflows for everything from audiobook chapters to in-app notifications, ensuring you can deliver high-quality synthetic speech with minimal code.

French Text to Speech API: Voice Selection & Integration Guide

Frequently asked questions

What is the maximum length for a single TTS request?

You can send up to 4096 characters per request. For longer texts, split the input into chunks of this size and join the resulting audio files. This ensures consistent processing and avoids payload limits while maintaining the integrity of the generated speech output.

Do voices support multiple languages natively?

Yes, every voice preset can speak all 23 supported languages. You do not need separate voices for different languages. Simply specify the target language in the request or let the system guess it based on the input script to generate the correct audio.

How does the API handle French text mixed with English?

The API detects the language from the script. If you mix languages, it may default to English for Latin script text. To ensure accurate French pronunciation, explicitly set the language parameter to fr in your request to guide the synthesis engine correctly.

Is the French voice female or male?

Both female and male voices support French. Female options include Aria, Nova, Luna, and Sage. Male options include Felix, Milo, Atlas, and Orion. All eight voices can speak any of the 23 supported languages. You can choose any preset based on your app’s tone, as all eight voices are capable of speaking French fluently.

Can I use the audio files commercially in my app?

Yes, you can use the generated audio files in your mobile applications. The input text is not used for training, and the audio is returned in the response without being stored. Each file contains an inaudible watermark marking it as synthetic speech for identification purposes.

What happens if the API request fails?

If synthesis fails, the charge is returned in full. You are not billed for failed requests. This ensures you only pay for successfully generated audio files, protecting your budget during integration testing or when handling edge cases in your mobile app workflow.

How many characters count towards the limit?

Spaces and punctuation count as characters. The limit is 4096 characters per request. Billing is based on the total character count, including all whitespace and symbols, so plan your text splitting accordingly to optimize costs and ensure smooth processing within your mobile application.

Do I need a subscription to use the API?

No, there are no monthly plans or subscriptions. You pay only for what you use, with pricing at $0.01 per 1,000 characters. Balance does not expire, and you can top up using USDT on the TRON network. This model suits variable usage in mobile apps.

Is there a limit on requests per minute?

Yes, the rate limit is 60 requests per minute per key. Before your first top-up, the limit is 10 requests per minute. This ensures fair usage and stable performance for your mobile app’s text-to-speech integration without requiring complex infrastructure management on your end.

How is the language detected if not specified?

If the language parameter is omitted, the API guesses the language from the script. Latin script text defaults to English. For accurate non-English Latin script output, you must explicitly set the language field to the appropriate ISO code to avoid incorrect synthesis.

Are the voices based on real human recordings?

Yes, the voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. This provides high-quality speech without the need for custom voice cloning.

Can I clone a specific French voice?

No, custom voice cloning is not offered. You can only choose from the eight preset voices: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion. These presets cover various tones and are available for all supported languages, including French, without additional cloning fees.

Does the API support SSML for French pronunciation?

No, the API does not support SSML. You cannot use phoneme dictionaries or custom pronunciation lexicons. Pronunciation is handled automatically by the engine based on the input text and specified language. For precise control, you may need to adjust the input text spelling.

How fast is the French text-to-speech generation?

Generation is roughly real-time on long texts, processing about 25 characters per second. Short phrases may take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue. Plan your app’s loading states accordingly for optimal user experience.

First request

curl

curl https://api.playk444.com/v1/audio/speech \
  -H "Authorization: Bearer $TTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
  --output hello.mp3

Why French TTS Matters for Mobile Apps

Sound wave illustration for: Why French TTS Matters for Mobile Apps

French text-to-speech integration enhances mobile apps by delivering natural, localized audio experiences that resonate with Francophone users. By converting on-screen text into clear speech, developers reduce cognitive load and improve content consumption for users who prefer listening over reading.

This approach supports accessibility standards, ensuring apps remain usable for visually impaired or dyslexic audiences. Additionally, audio playback often consumes less data than loading high-resolution images or video content, optimizing performance for users on limited mobile plans.

Localized User Experience

Localized User Experience

Native-sounding French voices create a more engaging and authentic interface. Users perceive the app as tailored to their linguistic preferences, increasing retention and satisfaction compared to generic robotic speech or static text alone.

  • Natural prosody and intonation
  • Support for 23 languages including French
  • Consistent brand voice across regions

Accessibility Compliance

Integrating TTS helps meet WCAG guidelines by providing alternative text-to-speech modalities. This is critical for users with visual impairments or reading difficulties, ensuring equal access to app content without requiring additional assistive hardware.

  • WCAG 2.1 AA compliance support
  • Screen reader alternative
  • Reduced reading fatigue

Reduced Data Usage

Audio files are generally smaller than high-resolution media assets. For content-heavy apps, serving short TTS clips instead of loading large images or videos can significantly lower bandwidth consumption, especially on slower mobile networks.

  • Smaller file sizes than video
  • Optimized for mobile data
  • Faster load times

Voice Catalog for French & Multilingual Apps

Sound wave illustration for: Voice Catalog for French & Multilingual Apps

VoiceScope provides eight distinct voice presets, each capable of speaking any of the 23 supported languages, including French. This flexibility allows mobile developers to maintain consistent audio quality across localized iOS and Android apps without managing multiple voice providers.

Aria: Warm Narrator

Aria offers a warm, clear narrative style ideal for audiobook chapters or educational content. Her tone remains consistent across long texts, ensuring listeners stay engaged without auditory fatigue.

Nova: Bright & Upbeat

Nova delivers a bright, energetic delivery suitable for promotional in-app notifications or quick updates. Her upbeat cadence helps capture user attention in short-form mobile interactions.

Luna: Soft Storyteller

Luna provides a soft, storytelling voice perfect for immersive experiences or language learning tools. Her gentle pace supports comprehension for non-native speakers using French content.

Technical Integration for iOS and Android

Sound wave illustration for: Technical Integration for iOS and Android

VoiceScope’s API mirrors the OpenAI speech endpoint structure, allowing iOS and Android developers to reuse existing TTS integration code with minimal changes. By simply updating the base_url to https://api.playk444.com/v1 and swapping the API key, your mobile app can immediately access high-quality synthesis without learning a new proprietary protocol.

This compatibility extends to official OpenAI SDKs, simplifying the development workflow for freelancers who need reliable text-to-speech functionality in both Swift and Kotlin environments without managing custom networking logic.

OpenAI SDK Compatibility

The endpoint POST https://api.playk444.com/v1/audio/speech accepts the same JSON body as OpenAI’s speech endpoint. This means you can use standard libraries like openai-python or the official iOS/Android wrappers by just pointing them to our base URL. Your existing authentication headers remain compatible with our Bearer tts-... token system.

  • Uses standard RESTful POST requests
  • Compatible with OpenAI SDKs via base_url override
  • No custom client libraries required

JSON Request Structure

Each request requires a JSON payload containing the model ("tts-1" or "tts-1-hd"), input (text up to 4096 characters), voice (e.g., "aria", "nova"), and response_format. You can optionally specify speed (0.25–4.0) and language (ISO code) to control output characteristics. For longer texts, split input into chunks, generate audio for each, and join the files in your app.

  • model: "tts-1" or "tts-1-hd"
  • input: String, max 4096 chars
  • voice: One of 8 presets
  • response_format: mp3, wav, opus, flac, pcm

Audio Format Options (MP3, WAV)

Choose mp3 for efficient bandwidth usage in mobile apps or wav for uncompressed, high-fidelity playback. Both formats are returned directly in the response body. MP3 is recommended for most mobile use cases due to smaller file sizes, while WAV suits scenarios requiring exact audio fidelity without decoding overhead.

  • MP3: Optimized for mobile data
  • WAV: Uncompressed quality
  • Direct response body delivery

Pricing Model for Freelancers

Sound wave illustration for: Pricing Model for Freelancers

VoiceScope uses a straightforward, character-based pricing model designed for mobile app developers who need predictable costs without overhead. You pay only for what you generate, with no hidden fees or minimum commitments.

This structure allows freelancers to scale usage up or down instantly based on app demand, ensuring that development budgets remain tight and transparent regardless of user growth.

$0.01 per 1,000 Characters

The standard rate is $10 per 1 million characters, which equals $0.01 per 1,000 characters. Spaces and punctuation count toward the total character count, so plan your text length accurately. This flat rate applies to both the tts-1 and tts-1-hd models, which share the same engine and pricing tier.

There are no volume discounts or tiered pricing structures to manage. Every request is billed uniformly, making it easy to calculate the cost of a specific feature or screen in your app.

  • - $10 per 1 million characters ($0.01 per 1,000 chars)
  • - Spaces and punctuation count as characters
  • - Same price for tts-1 and tts-1-hd

No Monthly Subscriptions

You do not need to commit to a monthly plan or reserve a minimum spend. Balance does not expire, so you can top up only when needed. This flexibility is ideal for freelancers managing multiple projects with varying TTS demands.

Payment is handled via USDT on the TRON network (TRC-20), with a minimum top-up of $5. Credits are added automatically within minutes, allowing you to start generating audio immediately without waiting for invoice processing.

  • - No monthly plans or seats
  • - Balance never expires
  • - Top-up via USDT (TRC-20), min $5

Free Failed Requests

If a synthesis request fails, the charge is returned in full. This ensures you are not penalized for transient network issues or invalid parameters. You only pay for successfully generated audio files.

This policy reduces risk when testing different voices or speeds during development, as you can experiment without worrying about accumulating costs for unusable outputs.

Supported Languages Beyond French

Sound wave illustration for: Supported Languages Beyond French

VoiceScope supports 23 languages, enabling mobile apps to reach global audiences without managing separate voice models. The catalog includes major languages like Arabic, Chinese, and Spanish, alongside regional variants such as Greek, Hebrew, and Turkish.

23 Languages Supported

The API covers high-demand markets including Swahili, Polish, and Russian. It also supports Swahili, Polish, Russian, and Dutch, ensuring broad compatibility for international mobile applications.

  • Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish

Automatic Language Detection

If the language field is omitted, the API guesses the language from the text script. Latin script text defaults to English, so for non-English Latin-script text you must set the language parameter.

Manual Language Override

For non-English Latin-script text, explicitly set the language parameter to the ISO code. This ensures accurate pronunciation and prevents the default English fallback from misinterpreting the input.

Supported Languages Beyond French

Performance and Speed Metrics

Sound wave illustration for: Performance and Speed Metrics

VoiceScope delivers audio at roughly real-time speeds, generating approximately 25 characters of text per second of audio. This efficiency ensures that long-form content, such as audiobook chapters or lengthy app narratives, processes without significant delay relative to playback duration.

For shorter phrases, generation typically takes a few seconds. Since requests are processed sequentially, heavy concurrent load may introduce queueing latency, but the system prioritizes accurate synthesis over raw throughput.

Real-Time Generation Speed

Text-to-speech synthesis runs at approximately real-time speed, producing about 25 characters of input text per second of audio. This ratio ensures that long texts, such as audiobook chapters or extensive in-app guidance, generate efficiently without disproportionate processing time relative to the resulting audio length.

Queue Handling Under Load

Requests are processed sequentially, meaning that under high concurrent load, a request may wait in a queue before synthesis begins. This linear processing model ensures consistent quality but may increase total latency during peak usage periods compared to parallelized systems.

Latency for Short Phrases

Short phrases typically take a few seconds to generate. While the per-character speed is high, the fixed overhead for initialization means that very brief inputs may feel slightly slower than real-time, though still suitable for interactive app notifications or quick UI feedback.

Data Privacy and Licensing

Sound wave illustration for: Data Privacy and Licensing

VoiceScope processes input text strictly to generate the requested audio file without retaining the content for model training. While request metadata such as timestamp, character count, and status are stored for billing accuracy, the actual textual data used in synthesis is not repurposed for improving the underlying engine.

No Training on Input Text

Your users’ private messages or app content remain isolated. Only the necessary metadata required for accurate billing and usage tracking is preserved. The raw input text is not ingested into our training datasets.

LibriTTS-R License

Our voice presets are derived from the LibriTTS-R dataset, licensed under CC BY 4.0. This foundation includes public-domain audiobook readings from LibriVox.

Synthetic Watermarking

Every audio file generated by VoiceScope contains an inaudible watermark. This digital signature clearly marks the output as synthetic speech, providing transparency for your end-users and helping to distinguish AI-generated audio from human recordings in your app.

Common Use Cases in Mobile Development

Sound wave illustration for: Common Use Cases in Mobile Development

Mobile apps benefit significantly from localized voice synthesis, reducing the need for expensive studio recordings while maintaining high user engagement. VoiceScope allows freelance developers to integrate high-quality French TTS directly into iOS and Android projects without managing complex infrastructure or licensing deals.

This approach ensures consistent audio quality across diverse content types, from casual reading to educational tools, while keeping data usage low and compliance with accessibility standards straightforward.

Audiobook Chapters

Developers can convert long-form text into audio files by splitting input into 4096-character chunks and joining the results. This method supports LibriVox-style public domain content, with each file carrying an inaudible watermark to identify synthetic speech.

Using voices like Aria or Luna provides a warm, clear narration style suitable for storytelling, ensuring listeners remain engaged without the cost of hiring voice actors for every new title.

  • Split texts into 4096-character parts for seamless audio joining
  • Supports public domain content with built-in synthetic watermarking
  • Ideal for Aria and Luna voices for narrative flow

In-App Notifications

Voice notifications provide immediate, accessible alerts for users who may not be looking at their screens. Short phrases typically take a few seconds to generate.

Formats like MP3 or WAV ensure broad compatibility across iOS and Android devices, while the character-based pricing keeps notification costs predictable and low, even for high-volume use cases.

  • Real-time generation ensures minimal delay for urgent alerts
  • MP3 and WAV formats ensure universal device compatibility

Language Learning Tools

Apps can offer pronunciation guides and reading practice using clear, instructional voices like Sage or neutral tones like Atlas. The ability to specify ISO language codes ensures accurate articulation, preventing the default English bias for Latin-script text.

Developers can generate endless practice phrases at a fraction of the cost of human recording, allowing for extensive content libraries that adapt to user progress without additional licensing fees.

  • Specify ISO codes for accurate non-English pronunciation
  • Use Sage for clear instructional guidance
  • Generate unlimited practice phrases with minimal overhead

Getting Started with VoiceScope

Sound wave illustration for: Getting Started with VoiceScope

Start by creating your account on the /get-api-key/ page using your Google identity. This method instantly activates your free trial credit, giving you 50,000 characters to test the API within seven days without committing funds upfront.

Once registered, navigate to your dashboard to generate your unique API key. This key authenticates your requests to the VoiceScope endpoint and allows you to monitor your character usage, balance, and request history in real-time.

Sign Up with Google

Visit the registration page and select the Google sign-in option. This process creates your account instantly and automatically credits $0.50 of trial balance. Note that email/password accounts do not receive this initial credit.

Generate API Key

From your dashboard, click to issue your API key. You will receive a token starting with "tts-". Store this securely, as it is the only credential needed for authentication. You can regenerate this key at any time, which immediately invalidates the previous one.

  • Keys are scoped to your account on playk444.com
  • Old keys stop working immediately upon regeneration

First Curl Request

Test your integration with a simple command. Replace YOUR_API_KEY with your generated token. This request uses the 'nova' voice to generate a short English phrase in MP3 format.

curl -X POST "https://api.playk444.com/v1/audio/speech" \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "tts-1", "input": "Hello from VoiceScope", "voice": "nova", "response_format": "mp3"}'

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. The underlying voice models are trained on data from LibriTTS-R, which aggregates public-domain audiobook readings from LibriVox.

Avatar of the voice aria
ariafemale
Avatar of the voice nova
novafemale
Avatar of the voice luna
lunafemale
Avatar of the voice sage
sagefemale
Avatar of the voice felix
felixmale
Avatar of the voice milo
milomale
Avatar of the voice atlas
atlasmale
Avatar of the voice orion
orionmale

23 languages

Any voice speaks any of these languages.

Sources and further reading

  1. Wikipedia
  2. openai-python on GitHub
  3. FFmpeg documentation
  4. Node.js documentation
  5. LibriTTS-R corpus (OpenSLR 141)
  6. LibriTTS-R paper (arXiv 2305.18802)
  7. Python documentation
  8. TRON developer hub: TRC-20
  9. Tether (USDT)
  10. LibriVox public-domain audiobooks
  11. curl documentation
  12. W3C WAI: making audio and video accessible

More on this site