// AI TOOLS · GROK VOICES

Grok Voices

xAI's upgraded voice generation for Grok — 21 new flagship voices (26 total), each natively multilingual across 25+ languages and cast for a specific job like support, characters, commentary, advertising, or education.

KompozyTurn one idea into a week of content — across every platform, published for you.
Get Started →

Last verified · 2026-07-13 · by Moe Ameen

What Grok Voices is

Grok Voices is xAI's voice-generation layer for Grok, and in early July 2026 it got a large expansion: 21 new flagship voices joined the original five, bringing the roster to 26. The pitch is naturalness and range — each voice is cast for a specific job rather than being one generic narrator, so the set spans support agents, character voices, commentary, advertising reads, and education. The new voices carry names like Carina, Orion, Luna, Iris, Altair, Zenith, Perseus, Helios, Kepler, Rigel, Cosmo, Celeste, Sirius, and Atlas, among others.

Every voice is natively multilingual, supporting all of Grok Voice's 25+ languages, so one voice keeps its character across languages instead of forcing you to pick a different narrator per market. The original five voices — Ara, Eve, Leo, Rex, and Sal — were retrained in the same update for more natural pacing, phrasing, and emphasis. Delivery is steerable through speech tags: you can insert pauses, laughter, breathing, whispering, and changes in volume, pitch, speed, and emphasis to shape how a line reads.

The voices are exposed to developers three ways: the realtime Voice Agent API (for live, conversational agents), the Text-to-Speech API (for generating spoken audio from text), and a Grok Voice Agent Builder for assembling agents without building the plumbing yourself. xAI also offers voice cloning from a reference clip up to 120 seconds, usable across the TTS and realtime APIs, currently limited to the United States.

Honest framing: Grok Voices is an audio engine, not a content app. It is very good at the voice layer — talking, narrating, cloning, speaking many languages — and priced per minute for apps that call it. What it does not do is produce a finished post: no captioned vertical video, no carousel, no blog or newsletter, no brand-voice governance, and nothing scheduled or published. Exact pricing and model details move with each release, so confirm current numbers in xAI's own docs before you build against them.

What you can make with it

  • Text-to-speech narration in any of 26 flagship voices, each cast for a role (support, character, commentary, ad read, education)
  • Multilingual voiceovers where the same voice keeps its character across 25+ languages for localizing one script
  • Real-time voice agents that listen and answer in natural speech via the Voice Agent API and Voice Agent Builder
  • Custom cloned voices from a reference clip up to 120 seconds, reusable across TTS and realtime
  • Steerable delivery — pauses, whispers, laughter, and pitch/speed/emphasis changes shaped with speech tags
  • The audio layer inside your own product — phone support, in-app voice, accessibility read-out

How Kompozy turns Grok Voices output into content

Grok Voices hands you a clean narration track; it does not hand you a video. That gap is exactly where Kompozy lives. Feed a script through the Grok TTS API in whichever flagship voice fits — an education voice for a tutorial, a punchy ad-read voice for a promo — and you get audio, not a post. Kompozy generates the moving picture that carries it: Persona Shorts and HeyGen avatar clips with a face-locked recurring identity, Clipped Shorts from longer footage, Listicle and Naturalistic Video over stock clips, and Marketing Shorts. Then it burns in word-synced captions, reframes to 9:16, 1:1, or 16:9 per feed, and schedules and publishes the finished set across nine social platforms plus blog and email from one queue with Autopilot and a per-post review pipeline.

The multilingual angle is where the pairing gets sharp. Because every Grok voice keeps its character across 25+ languages, you can voice the same script in five markets and keep one consistent narrator — then run each cut through Kompozy so the captions, framing, brand styling, and scheduling all match your look on every platform. And because Kompozy is bring-your-own-key on the Founding tier, a team already calling the Grok Voice API for narration keeps that audio pipeline while Kompozy owns the media rendering, the Persona Brief brand governance, and the publishing the voice engine was never built to do. Grok generates the voice; Kompozy turns it into a week of on-brand, captioned, scheduled video across every feed.

  1. Write the script and generate the voiceover with the Grok Text-to-Speech API in the flagship voice that fits the piece (support, character, commentary, ad, or education).
  2. For a localized set, render the same script in several of Grok Voice's 25+ languages using one voice so the narrator stays consistent.
  3. Bring the narration into Kompozy and generate the video that carries it — a Persona Short, a HeyGen avatar clip, a Listicle Video, or a Clipped Short.
  4. Let Kompozy burn in word-synced captions and reframe each cut to 9:16, 1:1, or 16:9 per platform.
  5. Schedule and publish the whole set across TikTok, Reels, Shorts, X, LinkedIn, and more from one queue with Autopilot.

Frequently asked questions

What are the new Grok Voices?

In early July 2026 xAI added 21 new flagship voices to Grok Voice, bringing the total to 26 (the original five plus the new set). Each voice is natively multilingual across 25+ languages and cast for a specific role — support, characters, commentary, advertising, or education. The original five voices (Ara, Eve, Leo, Rex, Sal) were also retrained for more natural pacing, phrasing, and emphasis.

How do I access Grok Voices?

Through xAI's APIs: the realtime Voice Agent API for live conversational agents, the Text-to-Speech API for generating spoken audio from text, and a Grok Voice Agent Builder for assembling agents. There is also voice cloning from a reference clip up to 120 seconds, usable across the TTS and realtime APIs.

How many languages do Grok Voices support?

Every flagship voice is natively multilingual across all of Grok Voice's 25+ languages, so a single voice keeps its character from one language to the next instead of requiring a different narrator per market.

Can Grok Voices create social media posts or videos?

No. Grok Voices generates the audio layer — narration, real-time voice agents, and cloned voices — but produces no captioned video, carousels, blogs, or scheduled posts. To turn a Grok voiceover into finished, on-brand content across platforms, run it through a content engine like Kompozy that generates the video, adds captions, and publishes it.

Related tools

  • Grok 4.5xAI's new flagship model — a fast, lower-cost reasoning model for coding, knowledge work, conversation, and multimodal understanding.
  • Grok Build 0.1xAI's fast agentic coding model — the engine behind the Grok Build CLI, built to write and ship software.
  • OpenAI Voice Models (2026 update)OpenAI's 2026 voice stack in the API — real-time speech-to-speech agents, live translation, streaming transcription, and steerable text-to-speech, all in one family.
  • SpeechifyA text-to-speech platform built around low-latency streaming voice — its Simba models turn any script into natural narration for reading, voiceover, and developer apps.
  • HeyGenAI avatar video platform that turns a text script into a talking-head video — in 175+ languages.

← All AI tools · Get started →