Podcast Text-to-Speech API: An Oddcast Alternative

By Kuwi Audio editorial team · Updated 2026-11-18

This page details how Kuwi Audio functions as a text-to-speech API for podcast publishers. We explain how to convert scripts into broadcast-quality audio using consistent voices like Aria or Atlas, with predictable pricing based on character count rather than complex tiered plans.

Readers will find specific integration steps for the OpenAI SDK, pricing breakdowns, and technical limits. This guide covers the entire workflow from account creation to your first audio generation, designed for teams managing episodic content production who need scalable, transparent API access.

Podcast Text-to-Speech API: An Oddcast Alternative

Frequently asked questions

How does Kuwi compare to Oddcast for podcast audio?

Kuwi offers a REST API with predictable pricing at $10 per million characters. Kuwi provides eight distinct voice presets and supports 23 languages out of the box. Unlike traditional TTS platforms, Kuwi requires no minimum spend or monthly subscriptions, making it ideal for independent podcast publishers seeking scalable, cost-effective audio generation.

What is the cost per character for text-to-speech?

The cost is $10 per 1 million characters, which equals $0.01 per 1,000 characters. Spaces and punctuation count toward the total character count. Failed requests are free, with the charge returned in full. There are no monthly plans, seats, or minimum spends. Balance does not expire, and you only pay for successful audio generation.

Can I use Kuwi for commercial podcast episodes?

Yes, Kuwi Audio is designed for podcast publishers converting scripts into broadcast-quality audio at scale. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings. Generated audio files are not stored, ensuring your input text is used solely for production. This model supports commercial episodic content production without additional licensing fees.

How many characters can I send in one request?

You can send up to 4,096 characters per request. For longer scripts, split the text into parts of this size and join the resulting audio files. This limit applies to both tts-1 and tts-1-hd models. Ensure your input text does not exceed this character count to avoid errors during synthesis.

Which voices are best for storytelling podcasts?

For storytelling, Luna offers a soft, storytelling tone, while Sage provides a calm, instructional voice. Aria is another option for warm, clear narration. These female presets are well-suited for narrative content. Male voices like Milo offer a friendly, conversational style. Choose based on your specific narrative style and audience engagement goals.

Is the API compatible with OpenAI SDKs?

Yes, the API endpoint shape matches the OpenAI speech endpoint. You can use official OpenAI SDKs by changing the base_url to https://api.playkuwi.com/v1 and using your site’s key. This allows seamless integration with existing codebases that already utilize OpenAI’s TTS structure, requiring minimal configuration changes for podcast automation workflows.

What audio formats are supported for export?

Kuwi supports MP3, WAV, OPUS, FLAC, and PCM formats. Select the format that best fits your podcast distribution platform. MP3 and OPUS are efficient for web delivery, while WAV and FLAC offer lossless quality. PCM is available for raw audio processing. Specify your preferred format in the request body during synthesis.

How long does audio generation take?

Generation takes roughly real time, about 25 characters of text per second of audio for long texts. Short phrases may take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue. Typical speed is sufficient for episodic content production without significant delays.

Do I need to specify language for each request?

If the language is not set, it is guessed from the script. Latin text defaults to English. For non-English Latin-script text, you should set the language parameter explicitly. Kuwi supports 23 languages, including Arabic, Chinese, French, German, Japanese, Spanish, and others. Specify the ISO code for accurate pronunciation.

Is there a free trial available for new accounts?

Yes, signing in with Google provides $0.50 of credit, valid for 7 days, which covers about 50,000 characters. Accounts created with email and password start without trial credit. This allows you to test the API and voice quality before committing to a top-up via USDT on the TRON network.

How are failed requests billed?

Failed requests are free. If synthesis fails, the charge is returned in full. You only pay for successful audio generation. Request metadata, including status, is kept for billing purposes. This ensures you are not charged for audio files that were not successfully produced due to API errors or other issues.

Can I generate audio for multiple episodes simultaneously?

Yes, you can send multiple requests concurrently. However, requests are processed one after another, so under load, a request may wait in a queue. Rate limits apply: 60 requests per minute per key. For large-scale production, consider splitting texts and managing queue times to ensure timely delivery of all episodes.

What happens if my API key is compromised?

You can issue a new API key at any time. The old key stops working immediately upon issuing the new one. Each account has one active key. This allows you to quickly revoke access if a key is compromised without needing to contact support or wait for manual intervention.

Is my input text used for training other models?

No, input text is used only to produce the audio and is not used for training. Generated audio is returned in the response and not stored. Request metadata is kept for billing. Your script content remains private and is not repurposed for model improvement or other commercial uses.

First request

curl

curl https://api.playkuwi.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 Podcast Publishers Choose Kuwi

Sound wave illustration for: Why Podcast Publishers Choose Kuwi

Podcast publishers require audio that sounds consistent across every episode. Kuwi provides eight distinct voice presets, from warm narrators to news-readers, ensuring consistent voice quality whether you are producing weekly shorts or long-form series.

Consistent Voice Quality

Consistent Voice Quality

Each voice is based on licensed recordings from LibriTTS-R and LibriVox, offering a professional broadcast standard without the variability of random selections. Every generated file carries an inaudible watermark, marking it clearly as synthetic speech for transparency.

Scalable API Integration

Our API endpoint mirrors the OpenAI structure, allowing you to swap in our base URL and use existing SDKs immediately. You can process up to 60 requests per minute, handling episodic batches efficiently without complex infrastructure changes.

Transparent Pricing Model

We charge $10 per million characters, with spaces and punctuation counting toward the total. Failed requests are free, ensuring you only pay for successful audio generation. There are no monthly subscriptions or hidden seat fees to manage.

Voice Library for Podcasting

Sound wave illustration for: Voice Library for Podcasting

Kuwi Audio offers eight distinct voice presets, each engineered for specific podcast formats. Female voices include Aria for warm narration, Luna for soft storytelling, Nova for upbeat segments, and Sage for instructional content. Male options feature Felix for energetic promos, Milo for conversational interviews, Atlas for neutral news delivery, and Orion for deep documentary tones.

All voices support 23 languages, allowing publishers to maintain consistent character voices across international editions. The API detects language automatically from the script, though explicit language codes are recommended for non-English Latin text to ensure accuracy.

Narrator and Storytelling Voices

Aria and Luna provide the foundation for narrative-driven episodes. Aria delivers clear, warm narration suitable for scripted intros or monologues. Luna offers a soft, storytelling quality ideal for serialized fiction or personal essay formats.

Sage serves as the neutral female option for instructional or educational podcasts. Its calm delivery ensures clarity for complex topics without distracting from the content. These voices are based on licensed LibriTTS-R recordings, ensuring high fidelity.

  • Aria: Warm, clear narrator
  • Luna: Soft, storytelling
  • Sage: Calm, instructional

Conversational and News Tones

Milo and Felix bring energy to commercial segments. Milo’s friendly, conversational style suits interview-style podcasts or casual updates. Felix’s energetic tone is optimized for promotional spots or dynamic show transitions. Atlas provides a neutral, news-style delivery for regular update episodes.

Orion’s deep, documentary voice adds gravitas to investigative or historical content. All eight voices share the same underlying engine, ensuring consistent audio quality regardless of the chosen persona. Each generated file includes an inaudible watermark for synthetic identification.

  • Milo: Friendly, conversational
  • Felix: Energetic, promo
  • Atlas: Neutral, news
  • Orion: Deep, documentary

Multilingual Support Options

The API supports 23 languages, including English, Spanish, French, German, Japanese, and Korean. Any voice can speak any supported language, allowing publishers to dub episodes without hiring separate voice talent. Arabic, Hindi, Russian, and Swahili are also available for global reach.

Language detection occurs automatically based on the script. For non-English Latin-script text, explicitly setting the language parameter prevents misidentification. Unsupported languages like Vietnamese or Thai are not available. This flexibility enables scalable multilingual production pipelines for episodic content.

  • 23 languages supported
  • Any voice speaks any language
  • Automatic language detection

Technical Specifications

Sound wave illustration for: Technical Specifications

The API supports five standard audio formats: MP3, WAV, Opus, FLAC, and PCM. Select the format that best matches your podcast delivery infrastructure and playback requirements.

Input text is limited to 4,096 characters per request. For longer scripts, split the text into chunks and join the resulting audio files. Generation speed averages roughly real-time, producing about 25 characters of text per second of audio.

Supported Audio Formats

Choose from MP3, WAV, Opus, FLAC, or PCM to suit your distribution needs.

MP3 offers efficient compression for web delivery, while WAV and FLAC provide lossless quality.

  • MP3
  • WAV
  • Opus
  • FLAC
  • PCM

Request Limits and Speed

Each request accepts up to 4,096 characters. Processing occurs sequentially, so requests may queue under high load.

Standard generation speed is approximately real-time. Short phrases may take a few seconds to complete.

Language Detection Logic

The system guesses the language from the script. Latin text defaults to English. For non-English Latin-script text, explicitly set the language parameter.

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

  • Arabic
  • Chinese
  • Danish
  • Dutch
  • English
  • Finnish

Integration Workflow

Sound wave illustration for: Integration Workflow

Kuwi Audio integrates seamlessly into existing podcast production pipelines by mirroring the OpenAI speech endpoint structure. Developers can route requests through the official OpenAI SDKs simply by adjusting the base_url to https://api.playkuwi.com/v1, ensuring minimal code changes for teams already familiar with standard TTS implementations.

OpenAI SDK Compatibility

The API accepts the same JSON body shape as OpenAI’s speech endpoint, supporting model, input, voice, response_format, and speed parameters. This compatibility allows publishers to use established client libraries without writing custom HTTP parsers.

  • Supports tts-1 and tts-1-hd models
  • Compatible with standard OpenAI Python/Node SDKs
  • Uses standard Bearer token authentication

cURL Quickstart Example

Testing requires a POST request to /v1/audio/speech with your API key. The following command demonstrates generating a 5-second clip using the 'nova' voice in MP3 format.

curl -X POST "https://api.playkuwi.com/v1/audio/speech" \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d "{\"model\": \"tts-1\", \"input\": \"Hello, welcome to the show.\", \"voice\": \"nova\", \"response_format\": \"mp3\"}" \ -o audio.mp3

Error Handling and Retries

Requests are processed sequentially, so under heavy load, responses may include HTTP 429 (Too Many Requests) if you exceed the 60 requests per minute limit. Failed synthesis attempts return a 5xx error with zero charges, allowing safe retries without double-billing.

Implement exponential backoff for rate limits. Since short phrases take a few seconds to generate, immediate retry loops are often unnecessary unless the error is specific to queue congestion.

  • Rate limit: 60 req/min per key
  • Failed requests are free
  • Sequential processing may cause temporary delays

Pricing Overview

Sound wave illustration for: Pricing Overview

Kuwi Audio charges $10 per 1 million characters, equating to $0.01 per 1,000 characters. Spaces and punctuation count toward this total. Failed requests are free, with charges returned in full if synthesis fails.

No monthly plans, seats, or minimum spend apply. Balances never expire, and there are no invoices or VAT. You pay only for the audio you actually receive.

Pay Per Character Rates

Billing is precise: every character, including spaces and punctuation, contributes to the $10 per million characters rate. Short phrases may take a few seconds to generate, while longer texts process at roughly real-time speed.

Free Trial Credits

Signing in with Google provides $0.50 in trial credit, valid for 7 days. This covers approximately 50,000 characters, allowing you to test voice quality before committing funds. Email-based accounts start without trial credit.

Balance does not expire, and there is no minimum spend requirement. You can use the trial to evaluate format options like MP3 or WAV.

USDT Top-Up Options

Funds are added via USDT on the TRON network (TRC-20). The minimum top-up is $5. Credits are automatically processed within minutes of transfer confirmation.

We do not accept credit cards, PayPal, or subscriptions. There are no enterprise contracts or invoices. Simply send the required USDT amount to your wallet address to restore your balance for ongoing podcast production.

  • Supported: USDT (TRC-20)
  • Minimum deposit: $5
  • Processing time: within minutes
Pricing Overview

Data Privacy and Licensing

Sound wave illustration for: Data Privacy and Licensing

Kuwi Audio processes input text exclusively to generate the requested audio file. We do not use your scripts to train our models or improve voice quality for other customers. Your content remains private and is not repurposed for machine learning datasets.

Input Text Usage Policy

The text you send via the API is used only for the immediate synthesis of the audio response. Once the audio is generated and returned, the input text is not stored for training purposes. This ensures your podcast scripts and proprietary content remain yours.

Audio Generation Storage

Generated audio files are returned directly in the API response and are not permanently stored on our servers. We retain only request metadata—such as timestamp, character count, and status—for billing and usage verification. This minimal data retention supports accurate invoicing without hoarding your audio assets.

Voice Source Licenses

Our voices are derived from licensed recordings of the LibriTTS-R dataset (CC BY 4.0) and public-domain readings from LibriVox. Each generated audio file contains an inaudible watermark identifying it as synthetic speech, ensuring transparency for broadcast and distribution platforms.

Common Use Cases

Sound wave illustration for: Common Use Cases

Kuwi Audio enables podcasters to convert scripts into broadcast-ready audio efficiently. The API handles the heavy lifting of synthesis, allowing creators to focus on content strategy while maintaining consistent voice quality across episodes.

With support for up to 4096 characters per request, publishers can generate substantial segments in single calls. The system processes requests sequentially, ensuring stable output even during high-volume episodic releases.

Audiobook Narration

Publishers can utilize the calm, instructional tone of Sage or the storytelling nuance of Luna to produce long-form content. By splitting texts into 4096-character chunks, the API joins audio files seamlessly, mimicking a continuous narration experience without manual editing.

  • Use Luna for soft, storytelling narratives
  • Split long scripts into 4096-character parts
  • Join outputs for continuous playback

Podcast Intro Generation

Generate consistent intro segments using energetic voices like Felix or friendly tones from Milo. Since short phrases take a few seconds to synthesize, you can quickly produce standard opening clips for episodic content without waiting for long processing times.

  • Use Felix for energetic promos
  • Use Milo for conversational intros
  • Fast synthesis for short clips

Multilingual Dubbing

Expand reach by dubbing episodes into 23 supported languages. Each voice can speak any supported language, allowing you to maintain brand voice consistency across markets. Specify the ISO language code to ensure accurate pronunciation for non-English scripts.

  • Support for 23 languages
  • Any voice can speak any language
  • Specify ISO code for accuracy

Comparison: Kuwi vs. Traditional TTS

Sound wave illustration for: Comparison: Kuwi vs. Traditional TTS

Traditional TTS solutions for podcasting often lock publishers into expensive monthly licenses or per-seat models, creating fixed overhead regardless of actual usage volume. Kuwi Audio replaces this with a pure pay-per-character structure, ensuring you only pay for the audio you actually generate.

Cost Efficiency Analysis

At $0.01 per 1,000 characters, costs scale linearly with your output. Unlike flat-rate enterprise contracts, this model eliminates waste for episodic shows with fluctuating episode lengths or experimental content, keeping production costs predictable and low.

Voice Realism Comparison

Our eight voices, derived from LibriTTS-R recordings, offer broadcast-ready clarity. While lacking granular emotion tags, the natural cadence of voices like Aria or Atlas provides a consistent, professional tone suitable for long-form narration.

API Flexibility Benefits

By adhering to the OpenAI endpoint standard, Kuwi integrates with existing workflows without requiring custom SDKs. Publishers can switch between models or voices via simple parameter changes, avoiding vendor lock-in and allowing easy integration with automation tools like Zapier or Make.

Getting Started

Sound wave illustration for: Getting Started

Begin by visiting the /get-api-key/ page to create your account. You can sign up using Google for an immediate free trial, or use an email and password combination.

Once registered, you will have access to your unique API key and initial credit balance, ready for integration into your podcast production workflow.

Account Creation Steps

Navigate to the registration page and select your preferred method: Google sign-in or standard email registration. Google accounts receive an immediate $0.50 trial credit valid for seven days, while email accounts start with a zero balance.

Verify your email address if you chose that method. Your account is now active and ready to accept requests on playkuwi.com.

API Key Generation

After logging in, locate your dashboard to view your active API key. Each account has one active key; issuing a new one immediately invalidates the previous one.

Copy the key starting with 'tts-' and store it securely. You will need this for all authentication headers in your API calls.

  • Keys are issued instantly upon account verification
  • Old keys stop working immediately upon new issuance
  • Store keys in environment variables for security

First Request Guide

Use the OpenAI SDK or a simple cURL command to send your first request. Point your base URL to https://api.playkuwi.com/v1 and set the Authorization header with your key.

Send a POST request to /v1/audio/speech with your script, choosing a voice like 'aria' or 'nova'. The system returns the audio file directly in the response.

  • Base URL: https://api.playkuwi.com/v1
  • Auth Header: Authorization: Bearer tts-...
  • Endpoint: POST /v1/audio/speech

Listen to the 8 voices

Each sample below was generated by this API from the same sentence.

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. FFmpeg documentation
  3. LibriVox public-domain audiobooks
  4. openai-node on GitHub
  5. Node.js documentation
  6. IETF RFC 7845: Ogg Opus
  7. LibriTTS-R paper (arXiv 2305.18802)
  8. ISO 639 language codes
  9. TRON developer hub: TRC-20
  10. openai-python on GitHub
  11. LibriTTS-R corpus (OpenSLR 141)
  12. MDN: Web Speech API

More on this site