Introducing 279 ElevenLabs Voices: The Biggest AI Voice Expansion for Podcasts

DIALOGUE now ships with 279 ElevenLabs voices — roughly 20 male and 20 female voices per language across English, Vietnamese, Japanese, Korean, Spanish, Chinese, and French. This is the single largest voice expansion in the product, replacing the previous 30-voice Gemini TTS library. Here is what changed, why, and how to use it.

From 30 to 279 Voices: What Changed

When DIALOGUE launched, you had about 30 Gemini TTS voices to choose from — a solid starting point, but limited in range. That meant fewer accent choices, fewer tone options, and occasional guesswork when picking hosts.

The new ElevenLabs library gives you 279 curated voices across all 7 languages. Each language gets roughly 20 male and 20 female options, so you can pair hosts that actually sound different — not two voices that blur together.

The expansion is not just about numbers. ElevenLabs voices bring more natural pacing, better emotional range, and clearer differentiation between speakers. For a two-host podcast format, that matters — the listener needs to know who is talking without the script announcing it every time.

Accent and Descriptive Labels: Browse Instead of Guess

Two new features make the enlarged catalog usable at scale:

Accent filter chips. The voice picker now supports 25+ accents — Australian, British, American, Kansai, Seoul, and more. Tap a chip and the list filters to voices matching that accent. Browsing by accent works across all languages, so you can find a Vietnamese voice with a northern or southern accent, or a Japanese voice with Tokyo or Kansai inflection.

Descriptive labels. Each voice is tagged with intuitive descriptors — calm, casual, confident, deep, chill, energetic, warm, authoritative, and others. These are the same labels ElevenLabs uses to categorize its shared voice library, not free-text tags. They surface directly in the voice picker, so you can scan for tone before even playing a preview.

Instant Previews from CDN

Voice previews used to require a round-trip to generate audio — a small delay that added up when browsing dozens of options. Every voice now has a pre-generated preview clip served from CDN. Tap a voice, hear it immediately. No waiting, no spinners.

The previews use a standard transcript across all voices, so comparisons are fair — same words, same pacing, different voice.

Why ElevenLabs? And What About Gemini?

ElevenLabs was chosen for three reasons:

1. Voice quality. ElevenLabs TTS consistently produces more natural-sounding speech with better prosody, especially for the conversational style a two-host podcast needs.

2. Accent diversity. The ElevenLabs shared voice library has far more accent variety across languages than Gemini TTS, which is critical for a multilingual product.

3. Speed. ElevenLabs audio generation is fast enough to serve the synthesis step without slowing down episode production.

Gemini has not been replaced — it is still the engine for AI research, topic grounding, and script generation. Only the text-to-speech provider changed. Gemini writes the podcast; ElevenLabs voices it.

Redesigned Voice Picker

The voice picker UI was rebuilt to support the larger catalog. Changes include:

• Accent filter chips at the top for one-tap browsing
• Usage-based ranking — voices you have used appear first
• Mobile-friendly layout that works on phone screens
• Descriptive labels visible in the list, not hidden behind a detail view

The goal was to make 279 voices feel manageable, not overwhelming. Filter by accent, scan the labels, play a preview, pick.

What This Means for Your Podcast

More voices mean more control over how your podcast sounds. If you produce content across languages, you can match host tones consistently — a warm, calm host pair for internal updates; a sharp, energetic pair for product launches. If you localize episodes, you can pick voice pairs that carry similar energy across languages rather than just matching gender.

The 279 voices are available at every pricing tier — no voice paywall. Start with 2 free episodes and hear the new voices yourself.

---

Try the new voice library. Create a podcast and browse all 279 voices with instant previews — free to start, no card required.