Gradium (@GradiumAI) header
Gradium (@GradiumAI) profile picture

The voice layer for agents. Real-time, accurate, natural voice APIs: TTS, STT, turn-taking & voice design. Devs: build → https://teapaw.org/t.co/BMTX2dOALw

Joined September 2025
Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)Photo by Gradium (@GradiumAI)
A new frontier for on-device speech. We're open-sourcing Phonon, a suite of TTS models under 50m parameters, setting a new standard for naturalness and accuracy across languages, with the broadest inference support to date. Phonon-31m is available under Apache 2.0. gradium.ai/phonon🧵
167
284
162
2,194
368,928
More and more open-source projects are building Phonon in. yovoice from @leemysw, an audio and music studio that runs fully locally, now uses it as one of its models for built-in voices. github.com/leemysw/yovoice Big thanks to everyone in the community building with Gradium. We love seeing where it ends up.
1
2
32
1,030
Our friends at @autonomous_labs got Lamp, their AI personal robot, talking with Phonon, our open-weight text-to-speech model that runs on-device. Ahead of Black Friday, they’re giving our community 20% off Lamp, starting today. Grab the code in the “Apps built” section of our Phonon page. Valid through Nov 30. gradium.ai/phonon
If you build with @GradiumAI, the real-time voice AI for agents and devices, Black Friday came early. 20% off Lamp, through Nov 30. Your code is waiting in the Gradium community. autonomous.ai/lamp
2
7
407
Welcome to our weekly community spotlight. Andre Louis, a visually impaired musician, built screen reader tools with Gradium Phonon for Windows, Mac, and iPhone. His iPhone app reached 1,300+ first-time App Store downloads in about two weeks: apps.apple.com/us/app/phonon…
1
2
3
12
1,544
Transcript: Builder Spotlight with Andre Louis Words in square brackets describe what is on screen. Quiet background music plays throughout. [Andre sits at his desk wearing headphones, with a music keyboard and a mixer in front of him. His screen reader reads text aloud at very high speed. On screen: "Can you keep up? This is how Andre reads a screen every day."] Andre: This is what I hear daily, right? I'm using it at this speed, and it can go even faster. [Title card: Builder Spotlight. Andre Louis.] Andre: Hi, so my name is Andre Louis. I'm actually a blind musician from London. I recently got into the AI coding scene in April, and I produced a lot of apps since then, all of them with accessibility in mind. [On screen: Andre Louis. A blind musician. From London. Coding with AI since April. Builds apps with accessibility in mind.] Andre: I thought to myself, the first thing I wanna try is can I make it usable via my screen reader? [On screen: Gradium Phonon, as the voice of your screen reader. Phonon is Gradium's open, on-device text-to-speech model.] [Andre's screen recording: the NVDA screen reader reads the gradium.ai home page in a Phonon voice.] Screen reader: Heading level two. Full suite of voice AI models to power your agents. Built to withstand the hard real world problems, latency, reliability, and scale. Andre: My NVDA add-on for Phonon, with its 102 voices across multiple languages, is less than 200 megabytes. [On screen: What it does. Gives NVDA a Phonon voice. 102 voices. Multiple languages. Under 200 MB. A short clip of the iPhone app, labelled "Also on iPhone".] Andre: And so that is usable because if you think of TTS lag like your screen refresh lag, right? Nobody wants to watch a ten-hertz screen. And that's the same kind of level of speed that we as blind people require on a daily basis. [On screen: Speech lag is like screen lag. Phonon keeps up. A diagram reads, left to right: a key press, Andre's Phonon app, Gradium Phonon, speech. Below it: Speech runs locally without a server or network connection, with smoother long-sentence phrasing, voice previews, and signed updates.] Andre: An English voice can speak German, a Portuguese voice can speak English. So you don't have to actually switch to another voice if you don't want to, and I think that's super powerful and super impressive. [On screen: What Andre liked about Phonon. An English voice can speak German. A Portuguese voice can speak English. Super powerful.] [Andre's screen reader reads his App Store statistics.] Screen reader: First-time downloads. Count: 1,306. [On screen: Phonon for iOS, about two weeks in. 1,306 first-time downloads. Free on the App Store. Android is next.] Andre: We're doing well, over 1,300. It hasn't been a long time this has been out, so to have that many downloads and they all be unique is really impressive. Andre: So it all starts with following a dream or trying to better the world that I live in in some small way. I want to make apps that people use daily, right? That they didn't know they missed until they had them. That is my whole ethos. Not just to make AI slop. That's furthest from my mind. And that's a goal I think that everybody should strive for and can achieve with a bit of care. [On screen: Advice to all builders. Follow a dream. Better your world in some small way. 1. Make what people will use daily. 2. Don't make AI slop. 3. Everybody can do it, with a bit of care.] [End card: Builder Spotlight. Andre Louis. Brought Gradium Phonon to screen readers. Gradium logo.]
2
181
The Gradium MCP is now in beta. Your coding agent can pick a voice and turn a script into speech in the same session where it writes your code. You sign in with your Gradium account, so you don’t need to create an API key. Get the setup prompt and read the full post: gradium.ai/blog/gradium-mcp-…
9
9
5
34
4,545
How do you describe a voice? We learned a lot by designing and testing 16,000+ voices while building our catalog, and from the prompts and feedback that teams building with Gradium share with us. The gap between the voice in your head and the one you get usually comes down to a few words. Our best prompts cover five parts: who is speaking, their role, accent, mood, and how they sound. We put it all in one post, with samples to compare: gradium.ai/blog/how-to-descr… What's the most unusual voice you've tried to design?
1
3
2
28
1,524
Ask a voice agent something it has to look up, and you’ll usually hear silence while it searches. Some time ago we built a voice agent with @KeenableAI that searches the live web mid-conversation and answers without the pause. Congrats to the Keenable team on launching 1.0 last week. Try our demo and ask it for this week’s latest AI news: gradium.keenable.ai
Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…
4
5
22
894
Last week we saw an iconic red bridge in two cities. Neil spoke at @Vapi_AI's VapiCon in San Francisco, while Alexandre joined @AgoraIO to talk about scaling voice AI and was on stage at @WebSummit in Lisbon. Next: → Nov 18, Helsinki: @neilzegh joins @mattturck live at @SlushHQ → Nov 28–29, Berlin: Big Berlin Hack with @techeurope_, register: gradium.link/big-berlin-hack
1
6
413
We can’t wait to see what you’ll build with Gradium. Registration for the Gradium × @TensorTonic contest is open: tensortonic.com/contests/gra…
Exciting announcement We're running a 30 day contest with @GradiumAI at TensorTonic. Build a voice app on Gradium's text-to-speech API. Voice is turning into the default way people talk to software, and this is a month to build something real on it with the team behind one of the strongest text-to-speech models available today. > 2M Gradium credits in prizes > A chance to interview for ML roles at Gradium > Internship interviews at TensorTonic > Exclusive merch for the top 5 > TensorTonic Plus for the top 10 Starts 20 Nov. Registration is live: tensortonic.com/contests
3
1
9
710
Welcome to our weekly community spotlight. Rosalie Wessels and Christian Zimmer won the Agentic AI Hack in Zurich with ∑cribe, a tutor for handwritten notes that you can talk to, powered by Gradium. Hear their story of building it with Gradium Text-to-Speech and Speech-to-Text models.
2
1
2
8
821
Describing a voice in writing is harder than it sounds. It’s much easier to say it out loud. That’s why we have an agent you can just talk to. Let’s say you need a British narrator. Add whatever details you have in mind, maybe he’s in his fifties with a warm, unhurried manner of speaking, and Gradium Voice Design brings him to life. Ask for “a bit lower” and it changes just that, keeping the same speaker. The code is open source: github.com/gradium-ai/gradbo…
2
2
1
22
915
Thanks @GoogleCloud_FR for having us at Google Startup Day! @neilzegh shared what drives us at Gradium: real-time voice models that turn every text interaction, including with your agents, into a natural, fast and pleasant conversation. Merci Google Cloud !
3
8
561
People don’t want another chatbot. They want a real conversation. German insurtech Jampai is making that happen with voice AI. Our voice Jette is in their ads, and it’s the same voice customers hear when they call. Watch our full Q&A with Michael Langer: youtu.be/jyzO0hIrMEg #VoiceAI
1
2
2
4
914
Some of our best conversations happen in person. Last week we hosted Voice Agents, After Hours with @covaldev in Paris. Thanks to everyone who came, it was such an insightful evening. More to come, here’s where to find us next ↓
2
3
11
579
Phonon ranks first on the @huggingface Open TTS Leaderboard with the lowest Time to First Audio.
A small, on-device TTS models is competing with models up to 100× their size, and taking first place on specific benchmarks! @GradiumAI's new Phonon model offers a concrete example. On the Open TTS Leaderboard, Gradium reports: - First place on intelligibility in German and Spanish. - Top-three intelligibility in every language it evaluated. - Leading CPU streaming results on time-to-first-audio. These results also show why it’s useful to compare TTS models on more than a single overall ranking. For the Open TTS Leaderboard, @EricBezzam separates several dimensions: - Intelligibility, measured through word/character error rates. - Generation throughput and time-to-first-audio. - Predicted naturalness scores. - Speaker similarity for voice cloning. Leading on intelligibility in a particular language is a specific result, and so is leading on CPU response latency; but neither establishes that a model is best across every use case. For on-device applications, Phonon’s combination of size, intelligibility, and CPU latency seems really relevant. For other (heavier?) workloads, language coverage, throughput, voice similarity, or delivery may be more important. Leaderboard: huggingface.co/spaces/hf-aud… Gradium’s post: huggingface.co/blog/phonon-t…
1
5
1
34
1,710
Welcome to our weekly community spotlight. Adem and Titouan aren’t engineers, but that didn’t stop them from winning the Paris AI Gaming Hack with Broskis, an arena game where everyday objects become weapons. Hear their story of building it with Gradium Voice Design and Text-to-Speech.
3
1
1
8
923