Add Spoken Feedback to Your App

Turn text into speech with Speakvora's API. Play confirmations and feedback messages to users with library phrases or live generation.

Short answer. Use Speakvora's /v1/speak endpoint to generate audio for feedback and confirmations. Call it from your server, get back an audio URL, and play it in your app. Library phrases are instant; new text is generated live.

How it works

When a user completes an action in your app—submitting a form, confirming a purchase, or triggering an alert—you can play a spoken message. Call /v1/speak from your backend with the text, voice, and language. The API returns an audio URL. Your app then plays that URL to the user.

Speakvora stores common phrases in a library. If your message is already there, you get the audio instantly. If it's new, the API generates it live and caches it for next time.

Use library phrases for speed

The quickest way to add feedback is to use fixed messages that are already in the library. Phrases like 'Order confirmed', 'Payment received', or 'File uploaded' return audio without generation delay.

  • 7,730 English phrases in the public catalog
  • Spanish, French, Portuguese, Hindi, Japanese, Chinese also available
  • No generation delay—instant playback
  • Billed at $3 per 1M characters

Generate dynamic text live

For personalized messages—'Hello, your order is ready'—use live generation. The API synthesizes new text on demand. Generation takes a few seconds; the audio is then cached.

  • Text limit: 1 to 500 characters per request
  • Supported languages: English, Spanish, French, Portuguese, Hindi, Japanese, Chinese
  • Same pricing: $3 per 1M characters
  • Queues under load

Server-side call example

POST /v1/speak

curl -X POST https://speakvora.com/api/v1/speak \
  -H "x-api-key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Order confirmed",
    "voice": "af_heart",
    "lang": "en"
  }'

# Response:
# {
#   "url": "https://...",
#   "voice": "af_heart",
#   "lang": "en",
#   "characters": 15,
#   "format": "m4a",
#   "live": false
# }

Play the audio in your app

Once you have the URL, play it on the client side. The URL is valid and cached for one year, so you can reuse it.

  • Audio format: M4A or MP3 (configurable)
  • URL is public and unguessable
  • Do not send personal data through catalog voices; use custom voices for sensitive content
  • For browser playback, use your backend to call /v1/speak, then return the URL to the client

Voices and languages

Choose from 12 live Kokoro voices for natural-sounding feedback. Each language has dedicated voices.

  • English: af_heart (Sanaya, female)
  • Spanish: ef_dora (Sofia, female), em_alex (Alex, male)
  • French: ff_siwis (Shaiza, female)
  • Portuguese: pf_dora (Dora, female), pm_alex (Alex, male)
  • Hindi: hf_alpha (Priya, female), hm_omega (Omega, male)
  • Japanese: jf_alpha (Alpha, female), jm_kumo (Kumo, male)
  • Chinese: zf_001 (Mandarin F1, female), zm_009 (Mandarin M1, male)

Pricing and free tier

Every account gets 100,000 free characters once, for any catalog-voice text. After that, you pay per character. Library hits and repeats are billed every time.

  • Catalog voices: $3 per 1M characters
  • Free allowance: 100,000 characters per account (one-time)
  • No monthly fees or minimums
  • Billing is per-character, not per request

Frequently asked questions

How long does it take to get audio back?

Library phrases return instantly. New text is generated live, then cached. Subsequent requests for the same text are instant.

Can I use this for personal data?

Catalog voices store audio at public URLs, so do not send personal data (names, account numbers, addresses) through them. Use a custom voice for sensitive content, which is private.

What if my message is longer than 500 characters?

Split it into multiple requests, each up to 500 characters. Call /v1/speak once per chunk and play them in sequence.

Do I need to store the audio myself?

No. Speakvora hosts the audio at a signed URL valid for one year. You just store and return the URL. The same text always returns the same URL.

Which languages support live generation?

English, Spanish, French, Portuguese, Hindi, Japanese, and Chinese. Other languages have library-only voices.

Related: API documentation, pricing, developer guides and more answers.