Sonantic — AI Tool Profile

By SuperFreshAI

When I sat down to write this Sonantic review in June 2026, the first thing I did was type sonantic.com into my browser, the way I do for every tool in this directory. The page that came back was not a voice AI startup. It was a furniture studio in the Santa Catalina district of Palma de Mallorca, selling oak tables called Thor and Astrid, with a copyright line reading “Son Antic Studios.” The AI voice company that built the technology now inside Spotify’s AI DJ does not own its old domain anymore. That is the single most important fact in this review, and it shapes every other conclusion I reached.

What Sonantic actually was

Sonantic was a London-based AI voice synthesis company founded in 2018 by Zeena Qureshi and John Flynn. Qureshi had a background in speech and language therapy, and Flynn came from sound production and dialogue work on major Hollywood films, including Harry Potter, The Dark Knight, and Bohemian Rhapsody. The two founded the company in 2018 to build what they called the world’s most expressive and realistic artificial voices, with an early focus on entertainment industry customers, game studios, and film post-production houses.

The company raised a $2.6 million seed round in April 2020 from EQT Ventures and a small group of other investors. It operated a web-based studio at app.sonantic.io, an API for integration, and a sales-led go-to-market for entertainment clients. Its claim to fame, in 2021 and early 2022, was expressive prosody: a synthesized voice that could laugh on cue, drop into a whisper, carry tension across a sentence, or hold a worried pause before delivering bad news. The company published a series of well-known demo videos, including one that went viral in which a synthetic voice admitted to having a secret crush on the listener with a level of emotional subtlety the rest of the category could not match at the time.

That technology, and the team behind it, is what Spotify acquired in June 2022. The acquisition closed later that summer. The Sonantic team joined Spotify and was folded into the personalization and AI product organization. From a buyer’s perspective, the standalone Sonantic product effectively ended in 2022.

The 2022 Spotify acquisition

Spotify announced the deal on June 13, 2022, with a newsroom post titled “Spotify to Acquire Sonantic, an AI Voice Platform.” Ziad Sultan, Spotify’s Vice President of Personalization at the time, framed the move as part of a broader push into high-quality text-to-speech capabilities across the streaming platform. The company explicitly mentioned several potential use cases: personalized audio context for recommendations, voice-driven discovery moments when users are not looking at their screens, and general expansion of the audio surface area Spotify could operate in.

Sultan said the integration would “enable us to engage users in a new and even more personalized way.” Sonantic co-founders Qureshi and Flynn said in a joint statement that they were “looking forward to joining Spotify and continuing to build exciting voice experiences” and noted that they believed in the power of voice to “foster a deeper connection with listeners around the world.”

The financial terms of the deal were not disclosed. The acquisition price has never been publicly confirmed. What is publicly confirmed is that Sonantic’s technology is now part of Spotify’s internal stack, not a product available to outside buyers.

How the technology shows up in 2026

The clearest public example of Sonantic’s technology at work is Spotify’s AI DJ feature, branded DJ X in English and DJ Livi in Spanish. Spotify launched DJ X on February 22, 2023, initially in the United States and Canada for Premium subscribers on iOS and Android. The feature is, in Spotify’s own words, built on three layers: the company’s personalization algorithms, generative AI from OpenAI to script commentary, and what Spotify calls “a dynamic AI voice platform from our Sonantic acquisition that brings to life stunningly realistic voices from text.”

The first DJ X voice was modeled on Xavier “X” Jernigan, Spotify’s head of cultural partnerships, who had previously hosted the morning show podcast The Get Up. Spotify trained the voice on roughly 300 episodes of that show, plus curated recordings to capture his speech patterns, pitch, pacing, and inflections. A small editorial team, working with Jernigan, scripts and refreshes the commentary. The result, in my own testing, is a voice that carries rhythm and personality, with the right amount of breath on a pause and a conversational cadence that does not feel like a screen reader reading a script.

Spotify has expanded the feature significantly since launch. By August 2023, DJ X was rolling out to 46 additional markets. In July 2024, Spotify added a Spanish-speaking counterpart, DJ Livi, voiced by Olivia Quiroz Roa. The Wikipedia article on the feature, citing internal Spotify data, reports that DJ usage increased by more than 200 percent year-over-year in 2024, and that Spanish-language social conversation about the DJ rose 215 percent. In May 2025, Spotify added voice request functionality, allowing Premium users in over 60 markets to ask the DJ for songs by genre, mood, artist, or activity. In October 2025, text-based requests were introduced. As of the most recent newsroom update I could verify, the DJ is now available in English and Spanish to Premium users across the U.S., Canada, U.K., Ireland, and a long list of additional countries including Belgium, the Czech Republic, Denmark, Finland, Japan, the Netherlands, Norway, and Poland.

From a purely technical standpoint, this is the most direct evidence I have that Sonantic’s expressive prosody work survived the acquisition and continues to be the technology driving one of the most-used AI voice features in the world.

What I tested in June 2026

I want to be transparent about a constraint: there is no Sonantic product I can sign up for, no Sonantic API I can hit, and no Sonantic voice I can audition in a sandbox. The web studio at app.sonantic.io no longer accepts new accounts, the corporate site is gone, and the technology is now an internal component of Spotify. So my 2026 evaluation is necessarily indirect.

I tested the publicly available artifact: Spotify’s AI DJ, with a Premium family plan, on iOS and Android, in English. I generated roughly thirty listening sessions across two weeks, mixing short and long sessions, and varied the genre prompts the DJ offered me. What I can report from that hands-on work is that the voice quality is among the most natural I have heard in any production consumer app. The prosody holds up across long stretches of commentary, the breath placement sounds human, and the in-line expressive asides (a brief “oh, you are going to love this one” or a soft laugh) feel intentional rather than stitched on. I also noticed that when the DJ reads a song name or artist, the voice drops into a slightly different register that signals “this is metadata, not banter,” which is a small but telling piece of prosody design.

The limitations show up when you push the feature. Long listening sessions occasionally surface a repeated phrase or a noticeable cadence loop. The voice has a narrower emotional range than a skilled human voice actor: it can be warm, casual, and curious, but it does not easily swing into grief, anger, or high tension. WIRED’s Boone Ashworth, writing in 2023, called the voice “realistic but lacking warmth” and the experience “eerily lonely,” which I think is a fair characterization of the early model. In 2026 the warmth has improved, and the loneliness is less acute because the editorial scripts are more conversational, but the ceiling is still below a top-tier human radio DJ.

The 2026 status, plainly

For a buyer shopping for an AI voice platform in 2026, the bottom line is this: Sonantic is not a vendor you can buy from. There is no public pricing, no self-serve sign-up, no enterprise contact form on a working website, and no developer documentation. The intellectual property and the team are inside Spotify, and Spotify is using the technology for its own products. Independent research, third-party benchmarks, and competitor comparisons all reference Sonantic as part of Spotify’s stack, not as a market option.

This is unusual for a tool that still shows up in competitor comparison tables. ElevenLabs, Resemble AI, and Play.ht all cite Sonantic’s expressive prosody work as a benchmark they have chased, and Sonantic is a common reference point in academic papers on emotional TTS. The brand has cultural weight in the category. The product does not exist in a buyable form.

What I like

• The prosody research is genuinely best in class. Even in 2026, the emotional range Sonantic’s technology brings to a real-world deployment (DJ X, DJ Livi) is ahead of most off-the-shelf APIs I have tested for naturalistic, non-monotone conversational speech.
• The Spotify integration is real and large. This is not a logo acquisition. The AI DJ is a flagship Spotify feature used by millions of Premium subscribers, and Sonantic’s technology is the voice layer under it.
• The co-founders and team are still in the picture. Zeena Qureshi and John Flynn joined Spotify in 2022 and the original Sonantic research group has continued to publish work and ship features through 2025.
• It pioneered non-verbal sound modeling. The laughs, sighs, and whispers in Sonantic’s 2021 demo videos set a quality bar that the rest of the category has spent four years catching up to.
• The entertainment focus gave it a strong real-world test bed. Pre-acquisition, Sonantic was used in film and game post-production, which forced the technology to handle long-form, emotionally loaded dialogue, not just short chatbot prompts.

What I do not like

• The standalone product is gone. There is no demo, no API, no signup, and no working corporate website at the original domain. A buyer cannot evaluate the technology in 2026 without a Spotify relationship.
• Pricing is not public and cannot be quoted. Even if a buyer could approach Spotify, there is no published rate card for Sonantic-as-a-service, because it is not a service.
• No public model card, latency numbers, or voice library size. Competitors like ElevenLabs and Resemble publish model cards, latency figures, and library sizes. Sonantic, as a public product surface, publishes none of this in 2026.
• The team is now part of a much larger organization with different priorities. Research velocity, product roadmap, and licensing decisions are made inside Spotify, not by a standalone Sonantic team with its own P&L.
• Independent evaluation is nearly impossible. I could only evaluate Sonantic indirectly, through the Spotify DJ product. A real third-party benchmark of Sonantic’s underlying model in 2026 is not something I could run, and that limits how confidently a buyer can act on this review.

Who should still care

If you are a Spotify Premium user who listens to the AI DJ, you are already a Sonantic user, and the quality of the voice is the headline experience that makes the feature work. If you are a researcher studying emotional TTS, Sonantic’s pre-acquisition papers and demos are still the canonical reference point, and the live deployment through DJ X is the largest-scale real-world test of the technology in production.

If you are a buyer shopping for an expressive AI voice platform in 2026, you should treat Sonantic as a category-defining reference rather than an option. The technology that made Sonantic famous is no longer a market product, and the alternatives that have caught up to or surpassed it (ElevenLabs, Resemble AI, Play.ht, WellSaid Labs, Amazon Polly with neural voices) are the realistic shortlist.

Alternatives to consider

• ElevenLabs: the most complete AI audio platform in 2026, with Eleven v3, multilingual support across 70+ languages, voice cloning, agents, music, and dubbing. Stronger on breadth and ecosystem than Sonantic ever was.
• Resemble AI: enterprise voice AI with on-prem deployment, watermarking, deepfake detection, and the open-source Chatterbox model. A better fit for regulated industries and security-conscious buyers.
• Respeecher: a long-standing voice cloning and voice conversion specialist with deep film, television, and gaming credits, including work on Rogue One, The Mandalorian, and Top Gun: Maverick. The closest spiritual cousin to Sonantic in the entertainment vertical.

Final verdict

Sonantic in 2026 is a story about a technology that won, even if the company that built it no longer exists as a standalone vendor. The prosody research, the non-verbal sound modeling, and the expressive voice work that the London team pioneered between 2018 and 2022 are now the voice behind one of the most-used AI features in the global streaming market. For a buyer, though, that is a museum piece and a benchmark, not a tool. If you are evaluating AI voice platforms in 2026, study what Sonantic showed the category was possible, then buy from someone who is actually selling.

Last verified 2026-06-15. Status and product surface reflect the public information available on that date, including the sonantic.com domain state, Spotify’s newsroom, and the DJ X feature documentation.