Voice Discoverability

Filtering by Parameters

One of the challenges with a catalogue of over 600 voices is the problem of choice. This presents is a discovery problem, very few users want to filter through all our voices!
So to deal with this we introduced parameters. Parameters are added and updated regularly.

List of Parameters:

  • provider (string) - Try one of: google, polly, azure, msnr, cerevoice
  • language (string) - Try with one of: english, spanish, french, german
  • accent (string) - Try with one of: american, british, neutral, portuguese/brazilian, american soft, mexican, australian, etc
  • gender (string) - Try with one of: male, female
  • ageBracket (string) - Try with one of: adult, child, senior
  • tags (string) - Try with one or more (separated by commas) of: steady, confident, balanced, informative, serious, instructional, slow, storytelling, calm, clear, deep, formal, sad, thin, fast, upbeat, fun, energetic, tense, very fast, flat, low pitched, high pitched, low-pitched, sing-y, cooperative, kind, stable, monotonous, neutral, responsible, business man, straight to the point, knowledgeable, focused, newscastery, newsreader, interviewer, reliable, friendly, welcoming, good for handing out information, slightly friendly
  • industryExamples (string) - Try with one or more of: fitness, business, commercial, fashion, travel, audiobook, faith, health industry, commercial, real estate, kids entertainment, games, customer service, education, storytelling, entertainment, kids, education audiobook
  • providerFullName (string) - Try with one of: 'amazon polly', 'google', 'microsoft azure', 'messner', 'retro', 'ibm', 'yandex', 'vocalid'
  • alias (string) - name of the voice
  • timePerformance (string) - The time performance of the voice. There are three categories: slow, medium, fast.
  • sectionCharacterLimit (string) - The maximum amount of characters that the voice can process per Script section.
    See more here


Listen to all our voices


Suggestions for Good UX in Voice Selection

API.audio’s Voice.list() is optimized to build for UIs for Voice Discoverability and Findability

Our library contains 600+ voices, in 60 different languages, accents, genders, ages and providers.We also offer specialized voices, different speaking styles, and niche voice styles. With all this to navigate through, making it easy for your users to choose a voice is complex, as there are hundreds of options, different brands, message fits and UX options.

To solve this, we have developed an extensive and well maintained tagging , labelling and ranking regime, spanning from voice tone, pitch, emotion, style, and suggested purpose, and much more. You can mix and match our filtering by purpose: based on content: e.g. Travel, Education, Fitness, by the gender of the speaker (M/F), by style: Fun, Upbeat, Energetic, by Language and Accent, and many many more.

Below there are some suggestions on how you can use our smart filtering system to build your own Frontend. We have curated 3 options for 3 main hypotheses; minimal, medium and large. Below we will present some UX styles and mockups, based on the hypotheses above.

Minimal Frontend

Curated shortlist for user based on input or job to do (eg. English script for female avatar). You can filter by VoiceName, Language/Accent, or even speaking style.


Option 1: By Voice name Option 2: By Language and Accent Option 3: By Speaking Style

Medium Frontend: Discovery Focused UX

Smart filtering can retrieve the list of voices for preview.

Option 1:


By Language/Accent, Gender and Age

Option 2:


By Language/Accent, Gender and Style

Option 3:


By Language/Accent and Keyword Search

Large: Parameter Search

Parameter based search for users to make the entire library searchable. This is the most comprehensible filtering system, and here are some combinations you can use:

Option 1:


By All filters available, including Purpose and Style

Option 2:


By Keyword/Name search

Option 3:


By Keyword/Name search, +Playground