What is Text-to-Speech API for Museum Audioguides?

By ChickenSubway editorial team · Updated 2026-11-18

This page explains how the VoxMuse Text-to-Speech API enables indie developers to generate professional museum audioguides programmatically. You will learn how to convert text into natural-sounding audio files using standard HTTP requests, select from eight licensed voices, and manage multi-language content for tourism experiences without complex infrastructure.

We cover the technical requirements, including JSON body structure and supported audio formats, alongside the straightforward pricing model based on character usage. Whether you are building an audioguide for a single museum or a multi-site tourism app, this guide details the integration workflow, data privacy terms, and known limitations to help you deploy reliable synthetic speech efficiently.

What is Text-to-Speech API for Museum Audioguides?

Frequently asked questions

What is a TTS API?

It is a cloud service where you send text and receive an audio file via a simple HTTP POST request. The endpoint follows the standard OpenAI speech schema, allowing you to integrate it easily using existing OpenAI SDKs or plain HTTP requests without needing custom infrastructure.

How does it work for museums?

Send up to 4096 characters of tour script to the API and receive synthesized audio. You can split longer narratives into parts, join the audio files, and deploy them to devices. The service supports 23 languages and eight distinct voice presets suitable for guides and narration.

Is it real human speech?

No, it is synthetic speech generated from licensed recordings of public-domain audiobook readings. Every audio file contains an inaudible watermark identifying it as synthetic. It is not a recording of a specific human performer but a realistic synthesis based on licensed voice data.

Can it speak multiple languages?

Yes, it supports 23 languages including English, Spanish, French, German, Japanese, and Arabic. Any voice can speak any supported language. If you do not specify the language, the system guesses it from the script, defaulting to English for Latin text.

Do I need to host servers?

No, it is a fully cloud-based API. You simply send requests to the endpoint and receive audio files. There is no infrastructure to manage, no servers to maintain, and no need for on-premise installations or data center choices.

Is it expensive for small tours?

Pricing is $0.01 per 1,000 characters. There are no monthly plans, seats, or minimum spends. You only pay for what you generate, and failed requests are free. A free trial provides $0.50 credit for initial testing.

Can I change the voice mid-tour?

Yes, you can select any of the eight voice presets for each request. You can switch between voices like Aria for narration or Felix for promos within the same tour by specifying the voice parameter in each API call.

Does it work offline?

No, it requires an internet connection to send text to the cloud API and receive the audio file. There is no offline mode, local SDK, or embedded player included. You must download the generated audio to play it offline.

Is the audio copyrighted?

The voices are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. The generated audio files are not stored by the service and are returned directly in the response. You can use the output for your tours without additional licensing.

Can I use it for podcasts?

Yes, you can generate audio files for podcasts by sending text to the API. The service supports standard formats like mp3 and wav. You can control speed and voice to match your podcast style, paying only for the characters generated.

How long does generation take?

Generation is roughly real-time, producing about 25 characters of text per second of audio. Short phrases take a few seconds, while longer texts may take longer. Requests are processed sequentially, so high load may cause slight queue delays.

What if the internet is slow?

The API processes requests sequentially. If your connection is slow, the request transmission may delay, but generation speed remains consistent. Under high load, requests may wait in a queue. Failed requests are free: if synthesis fails, the charge is returned in full.

Do I need to code?

Minimal coding is needed to send HTTP requests. You can use the official OpenAI SDK by changing the base URL, or use standard HTTP libraries. A ready curl example is provided on the dashboard for quick integration without complex development.

Is there a free tier?

Yes, signing in with Google provides a $0.50 trial credit valid for 7 days. Email-based accounts start without trial credit. You can top up with USDT on TRON for further usage. There are no ongoing free tiers, just pay-as-you-go.

Definition of TTS API

Sound wave illustration for: Definition of TTS API

Text-to-Speech APIs convert written content into audible audio files via a simple cloud endpoint. This technology enables developers to transform static text into natural-sounding speech without managing complex infrastructure or hardware.

Text to audio conversion

Send text up to 4,096 characters and receive an audio file in formats like MP3 or WAV. The process handles character counting accurately, including spaces and punctuation, ensuring precise billing based on actual usage.

  • Converts up to 4,096 characters per request
  • Supports MP3, WAV, OPUS, FLAC, and PCM
  • Accurate character counting for billing

Cloud-based processing

All computation happens on remote servers, eliminating the need for local installation. You only need an internet connection to send requests and receive audio files, making it ideal for mobile audioguide applications.

Programmatic access

Access the API via standard HTTP POST requests. The endpoint follows the OpenAI speech format, allowing you to use existing SDKs or simple cURL commands to integrate TTS directly into your tourism app code.

Authentication is handled via a Bearer token, ensuring secure access to your generated audio assets. This programmatic approach allows for automated content updates and dynamic audioguide generation based on user location or selection.

  • Standard HTTP POST endpoint
  • OpenAI SDK compatible
  • Bearer token authentication

Use in Tourism

Sound wave illustration for: Use in Tourism

A TTS API transforms static museum text into professional audio guides. Instead of recording every exhibit description manually, you send JSON requests to generate natural-sounding narration. This lets indie developers produce audioguide content efficiently, turning descriptive text into listenable experiences without hiring voice actors for every update.

The API supports 23 languages, allowing museums to offer guides in Arabic, Chinese, English, French, and more. Any voice can speak any supported language, ensuring consistent quality across translations. This makes it ideal for international tourists who need clear, accurate audio in their native tongue.

Audioguide content generation

Convert exhibit descriptions directly into audio files. Send text via a simple HTTP POST request and receive MP3 or WAV output. This workflow replaces expensive studio recordings, enabling rapid production of guide content for hundreds of exhibits.

Each request handles up to 4096 characters. For longer plaques, split the text and join the resulting audio files. The system processes requests sequentially, so under load a request may wait in a queue.

  • Convert text to MP3, WAV, OPUS, FLAC, or PCM
  • Handle up to 4096 characters per request
  • Join split audio files for long exhibits

Multi-language support

Support 23 languages including English, Spanish, German, and Japanese. Any voice preset can speak any supported language, providing flexibility in guide personality regardless of language.

If the language is not specified, the API guesses from the script. Latin text defaults to English, so explicitly set the language code for non-English Latin-script text to ensure accuracy.

  • 23 supported languages
  • Any voice can speak any language
  • Explicit language codes for non-English Latin text

Dynamic content updates

Update guide content instantly without re-recording. Change exhibit text in the database, and the API generates new audio on demand. This is ideal for temporary exhibitions or rotating displays.

No need to manage large audio libraries. Store text and generate audio when needed, reducing storage costs and ensuring the latest information is always available to visitors.

  • Instant content updates via API
  • Ideal for temporary exhibitions
  • Reduced audio storage costs

Technical Requirements

Sound wave illustration for: Technical Requirements

Send a POST request to https://api.getchickensubway.com/v1/audio/speech. The API accepts standard JSON payloads, allowing you to use official OpenAI SDKs by simply updating the base_url. This compatibility lets indie developers integrate museum audio content using familiar tools without learning a proprietary protocol.

HTTP requests

Authentication requires a Bearer token in the header: Authorization: Bearer tts-.... Each account holds one active key, which can be regenerated instantly if compromised. Requests are processed sequentially, so complex audioguide generation may queue under heavy load. Typical speed is roughly real time on long texts.

  • Method: POST
  • Endpoint: /v1/audio/speech
  • Auth: Bearer token

JSON body structure

The request body includes model ("tts-1" or "tts-1-hd"), input (up to 4096 characters), voice (one of 8 presets), response_format, and optional speed (0.25–4.0). For longer museum tours, split text into chunks and join the resulting audio files programmatically.

Audio format output

Choose between mp3, wav, opus, flac, or pcm via response_format. MP3 is recommended for efficient storage in mobile audioguide apps, while WAV offers lossless quality for archival purposes. The API returns the audio file directly in the response body.

Voice Selection

Sound wave illustration for: Voice Selection

Choose from eight distinct voice presets, each designed for a specific narrative role. Options range from warm, clear narrators to calm, instructional guides, ensuring your museum content sounds professional and consistent.

8 preset voices

The API offers eight distinct voice presets: Aria, Nova, Luna, and Sage for female tones, and Felix, Milo, Atlas, and Orion for male tones. Each voice has a unique character, from energetic to deep and documentary-style.

  • Aria: Warm, clear narrator
  • Nova: Bright, upbeat
  • Luna: Soft, storytelling
  • Sage: Calm, instructional
  • Felix: Energetic, promo
  • Milo: Friendly, conversational
  • Atlas: Neutral, news
  • Orion: Deep, documentary

Gender and tone variety

Select voices that match your exhibition’s mood. Whether you need a friendly conversational tone for interactive displays or a neutral news style for historical facts, the variety covers standard audioguide requirements without needing custom voice cloning.

Matching guide style

Ensure the voice aligns with your brand. Use Sage for instructional museum tours or Orion for deep-dive historical documentaries. All voices support 23 languages, allowing a single voice to maintain consistency across international exhibits.

Language Support

Sound wave illustration for: Language Support

Our API supports 23 languages, enabling museums to serve diverse international visitors without managing separate voice assets. Each voice can speak any supported language, ensuring consistent audio quality across languages.

Auto-detection works for Latin-script text, defaulting to English. For non-English Latin text, set the language explicitly to avoid misidentification. Unsupported languages include Vietnamese, Thai, and Ukrainian.

23 supported languages

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

Any voice can speak any of these languages, allowing you to maintain a consistent guide persona across different linguistic audiences.

Auto-detection limits

If the language field is omitted, the system guesses from the script. Latin text defaults to English.

This works for English but may fail for other Latin-script languages. It does not support Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali and all others not in the list.

Explicit language codes

Set the language parameter with an ISO code for non-English Latin text to ensure accurate synthesis.

This prevents misidentification and ensures the correct pronunciation for your museum's content.

Cost Structure

Sound wave illustration for: Cost Structure

Our pricing is straightforward and based entirely on character count. You pay $10 per million characters, which equals $0.01 per 1,000 characters. Spaces and punctuation count toward this total. If a synthesis request fails, the charge is returned in full, ensuring you only pay for successful audio generation.

There is no monthly subscription fee, no required minimum spend, and no enterprise contracts. Accounts created with email and password start without trial credit. Signing in with Google provides $0.50 valid for seven days. Top-ups are processed via USDT (TRC-20) with a $5 minimum, credited automatically within minutes.

Pay per character

The cost is calculated per character of input text. At $10 per million characters, long audioguide scripts remain affordable. Requests are processed sequentially, so under load, generation may take slightly longer, but the character-based billing remains consistent.

  • $0.01 per 1,000 characters
  • Spaces and punctuation count
  • Failed requests are free

No minimum spend

You can top up exactly what you need. There are no seats, no monthly plans, and no balance expiration. The balance does not expire, allowing you to pause usage between museum seasons without losing credits.

Trial credit available

Signing in with Google provides $0.50 of credit (approx. 50,000 characters) valid for 7 days. Email/password accounts do not receive this trial. This allows indie developers to test integration with standard tools before committing to USDT top-ups.

Cost Structure

Data Privacy

Sound wave illustration for: Data Privacy

Your input text is used strictly to generate the audio file and is not added to training datasets. This ensures that your specific tour scripts or historical narratives remain private and are not repurposed for model improvement.

Generated audio files are returned directly in the API response and are not stored on our servers after delivery. This minimizes data retention and reduces the risk of accidental exposure of your proprietary content.

Request metadata, including timestamp, character count, and status, is retained solely for accurate billing and usage tracking. No other personal or content data is kept long-term.

Input text not trained

Your input text is used strictly to generate the audio file and is not added to training datasets. This ensures that your specific tour scripts or historical narratives remain private and are not repurposed for model improvement.

No stored audio

Generated audio files are returned directly in the API response and are not stored on our servers after delivery. This minimizes data retention and reduces the risk of accidental exposure of your proprietary content.

Metadata for billing

Request metadata, including timestamp, character count, and status, is retained solely for accurate billing and usage tracking. No other personal or content data is kept long-term.

Deployment Model

Sound wave illustration for: Deployment Model

Our Text-to-Speech API operates entirely in the cloud, eliminating the need for local infrastructure. You send text via standard HTTP requests and receive audio files directly, allowing indie developers to integrate audioguide narration into tourism apps without managing hardware.

Cloud API

Processing happens on our servers. You interact with the API endpoint POST https://api.getchickensubway.com/v1/audio/speech using JSON bodies compatible with standard SDKs. This ensures consistent performance regardless of your device's processing power.

No server maintenance

Since the engine is cloud-based, you do not handle updates, patches, or server uptime. Requests are processed one after another, so under load a request may wait in a queue.

Scalable usage

Usage scales with your token limits. With 60 requests per minute per key, you can generate content for large tourist groups or high-traffic museum exhibits without manual intervention. If demand spikes, requests queue automatically.

Quality Assurance

Sound wave illustration for: Quality Assurance

Quality assurance for museum audioguides relies on transparent sourcing and consistent technical standards. Voices are based on licensed recordings: LibriTTS-R (CC BY 4.0), from LibriVox public-domain audiobook readings.

Every generated audio file contains an inaudible synthetic watermark. This allows institutions to verify that the audio was produced by our API, aiding in content management and rights tracking.

The output remains consistent across generations because the same engine processes each request. This reliability ensures that your audioguide maintains a uniform tone throughout the museum experience.

Licensed voice samples

Our voices are based on licensed recordings from the LibriTTS-R dataset (CC BY 4.0) and public-domain LibriVox audiobook readings. This foundation provides a natural, human-like quality without the need for complex voice design.

Since the source material is licensed, you avoid unexpected copyright issues when deploying audioguides in public spaces or digital apps.

  • Source: LibriTTS-R (CC BY 4.0) and LibriVox public-domain recordings

Synthetic watermark

Each audio file generated by the API carries an inaudible watermark. This digital signature marks the file as synthetic speech produced by our specific engine.

This feature helps museum curators and developers identify API-generated content during audits or when integrating multiple audio sources into a single tour experience.

Consistent output

Because requests are processed through the same underlying engine, the voice characteristics remain stable across different text inputs. This consistency is crucial for maintaining a professional tone throughout long audioguide sequences.

Even when switching between voices or languages, the base audio quality and processing standards remain uniform, ensuring a predictable user experience for your visitors.

Integration Speed

Sound wave illustration for: Integration Speed

Integration is straightforward because the API follows the same request shape as the OpenAI speech endpoint. You can use official OpenAI SDKs by simply changing the base URL to https://api.getchickensubway.com/v1. This allows indie developers to reuse existing codebases with minimal adjustments, speeding up the development of audioguide applications.

For quick testing, a ready curl example is available on the dashboard. The standard endpoints handle voice selection, format output, and speed adjustments in a single JSON body, ensuring that integrating text-to-audio conversion into your tourism app takes only minutes rather than days.

OpenAI SDK compatibility

The API mirrors the OpenAI speech endpoint structure. By updating the base URL to https://api.getchickensubway.com/v1, you can use any standard OpenAI SDK. This means you can use the official OpenAI SDKs by changing the base_url.

Your existing code for handling authentication and JSON bodies works immediately. This reduces integration time significantly for developers already familiar with the OpenAI ecosystem.

  • Use standard OpenAI SDKs
  • Change base_url to api.getchickensubway.com
  • Reuse existing JSON structures

Quick cURL setup

The dashboard provides a ready-to-use curl example. You can copy and paste this command to verify your API key and balance instantly. This is ideal for rapid prototyping before building the full application logic.

No complex configuration files are required. The endpoint accepts standard parameters, allowing you to generate your first audio file in under a minute.

  • Copy-paste ready examples
  • Instant verification of keys
  • Minimal setup steps

Standard endpoints

The API uses standard HTTP methods. POST to /v1/audio/speech for generation, GET /v1/voices for metadata. This predictability makes debugging easier and ensures compatibility with most modern development frameworks.

You do not need to manage custom endpoints for different features. All voices, models, and formats are accessible through the same consistent structure, simplifying your codebase.

  • POST /v1/audio/speech
  • GET /v1/voices
  • GET /v1/models

Limitations

Sound wave illustration for: Limitations

Our API is designed for reliable, on-demand synthesis rather than real-time streaming. You send a text request and receive the complete audio file once processing finishes. This ensures consistent, high-quality output without the complexity of managing WebSocket connections or partial audio buffers, which suits pre-loaded museum guide apps well.

The system processes requests sequentially, so under heavy load, longer texts may wait in a queue. There is no official SDK; we rely on standard HTTP and OpenAI-compatible libraries. Support is provided via email only, with no phone or live chat options available.

No streaming

The API returns the full audio file upon completion. There is no WebSocket support for real-time partial audio playback. This model is ideal for museum audioguides where audio files are pre-generated or cached, ensuring stable playback without buffering issues during the user's tour.

No custom cloning

We offer eight preset voices based on licensed recordings but do not provide custom voice cloning. You cannot upload your own voice samples to create a unique guide. If you need a specific voice, you must choose from the available presets like Aria, Nova, or Atlas.

  • No voice cloning
  • No custom voice upload
  • Fixed preset library

No SSML support

The API does not support SSML (Speech Synthesis Markup Language). You cannot control precise pauses, emphasis, or pronunciation via XML tags. Speed adjustments are limited to the speed parameter (0.25–4.0), and pronunciation relies on the model's internal understanding of the text.

  • No SSML tags
  • No manual pause control
  • Basic speed adjustment only

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. We build immersive audio stories for independent creators who want to bring museums to life through sound.

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. curl documentation
  3. LibriVox public-domain audiobooks
  4. OpenAI API reference: create speech
  5. TRON developer hub: TRC-20
  6. Tether (USDT)
  7. W3C WAI: making audio and video accessible
  8. Node.js documentation
  9. ISO 639 language codes
  10. MDN: Web Speech API
  11. openai-python on GitHub
  12. Creative Commons Attribution 4.0

More on this site