Swedish or Catalan voice actors, paid AI voice project opportunity We are looking for natural, expressive speech to support text-to-speech R&D. Selected talent would record between 2–10 finished hours, with session work paid at $80/hour. A few important details upfront: 🇸🇪 Swedish or Catalan speakers 💰 $80/hour for recording sessions 🎙️ 2–10 finished hours if selected 🔞 18+ / local age of majority 🚫 No union/guild affiliation that prevents participation 🌎 Talent cannot be located in California, Illinois, Nevada, Texas, Washington or Canada ⏰ Round-one auditions are being reviewed September 2 If you're interested, you can audition here: form.voices.com/t/17BCNuXcJ2… Happy to answer general questions about the opportunity below. #swedish #vojobs #jobs

So excited to welcome @Voices as a new Exhibitor to ICASSP 2026! Find out more about what they do at hubs.la/Q049HTgF0 or visit them at Stand #E13. See you in Barcelona!

The future of character in AI starts here. We’ve launched the world’s first character audio dataset, purpose-built for training voice AI. • 450+ character types • 35 emotions • 19 tones • Recorded by professional voice actors Every file is ethically sourced, metadata-rich, and ready to train, so developers can build AI voices that are truly expressive, diverse, and human-first. Explore what’s possible: voices.com/landing/character…

Voice AI just leveled up. We’re introducing the first-ever multi-character audio dataset—built with over 450 unique character voices, recorded by professional talent. Voice data is scarce, but demand is soaring. Training AI on ethically sourced, richly expressive voices means more natural, human-sounding results. From drill sergeants to wizards, this dataset gives developers the creative range they’ve been waiting for. Without compromising on ethics. Explore the dataset at the link below. voices.com/landing/character…

How do you get over 28,000 high-quality video clips in 22+ languages for an AI training project? We did it with 3,200 contributors from around the world. Check out our latest case study to see how we delivered on this massive data initiative. Link is below. voices.com/client/stories/ca…

Ethical AI doesn’t have to be complicated. We made it simple. Our Free Ethical AI Checklist lays out everything you need to keep your AI voice projects transparent, fair, and responsible—from disclosure and consent to bias prevention and security. It takes two seconds to download, and it’ll save you a lot of second-guessing down the road. Seriously, get the checklist, keep it handy, and know you’re covered. Download the free Ethical AI Checklist today. Link is below. voices.com/landing/ethical-a…

Find your Voice Over. Find your Voice AI. Find your Voice Data. Find your voice at Voices.com.

We make it easier than ever for talent to earn more, without more work. Now, clients can renew broadcast licenses directly on the platform. That means no more tracking deadlines or chasing follow-ups. We handle it all, and you get paid again for work you’ve already done. And that’s just one of the new ways we’re helping talent unlock additional income streams. 👉 Read the full update here: voices.com/blog/new-income-p…

High-quality AI starts with high-quality data. That’s why every audio file we deliver goes through two layers of testing, automated for technical precision, and human for all the stuff that robots don't get. Because you wouldn’t launch your product without testing it at least twice. Neither would we. 👉 Follow us for more on building better AI with better voice data.

Your AI voice is only as good as the data you train it on. It’s not just about clean audio, it’s about voices that are emotionally, tonally, rhythmically, and regionally diverse. And you can’t fake that. That’s why the best AI voice models are built on ethically sourced datasets, where every voice has given specific consent, been properly compensated, and retains control over their work. If you care about the quality of your AI, you need to care about the quality of your data.

We're confident. And you can be, too. When you're evaluating partners, you need to be sure they can deliver. That's why we're happy to provide a complete sample of our work. Our sample package includes everything you need to make an informed decision: audio, speaker tags, transcripts, and emotional cues. This allows you to evaluate the full package and get the internal alignment you need to move forward quickly. We're not just giving you a preview, we're making a promise. We're ready, capable, and committed to delivering the high-quality deliverables your model deserves. Ready to hear the difference? Send us a message to request your sample today.

Voice AI is having its gold rush moment. But as our CEO Jay O'Connor writes in his new @fastcompany piece, not all that glitters is ethical, legal, or even usable. Voice AI is only as good as the data behind it. And if that data is scraped, low-quality, or lacks consent... we call that fool’s gold. Jay lays out a clear case for why high-quality, licensed, and human voice data is the only way forward for tech that’s ethical and effective. Read the full article at the link: fastcompany.com/91379128/the… #VoiceAI #EthicalAI #AIethics #VoiceData #GenerativeAI

Your AI voice still sounds like a robot? It’s not your fault, it’s the training data. Wanna fix it? Use real voices from real people who were paid properly and knew what they were doing. Follow us if you want your AI to sound human, too.

Think your voice dataset is ethical? Here’s how to actually know for sure. ✅ Clear, informed consent ✅ Fair pay for contributors ✅ Transparent use of every voice If you can’t check all those boxes, your brand and your customers’ trust are at risk. Ethics isn’t a nice to have in AI. It’s the foundation. If you agree, hit follow.

If someone is claiming that their voice dataset is ethical and high quality, but they're asking anyone with an iPhone to record data without mentioning pay... 🚩🚩

Most people think AI voice quality is all about the voice itself. It’s not. The real power is in the metadata. Not just the basics. We’re talking detailed tags for emotion, mic setup, background noise, pacing, speaker identity, and more. This is what gives an AI voice real-world context. Without it, your model is basically guessing. Good metadata means your model learns with intention. Bad metadata, or none at all, means you’re just scraping noise. And scraped datasets from the internet? They’re not just ethically questionable. They sound worse. We include rich, precise metadata with every audio file we deliver. Because when it comes to building voice AI that actually works, context isn’t optional. Want to see what that looks like? Follow along.

“Voices has been an incredible partner for our production team. From ads to creative campaigns, finding the right voice talent is easy and stress-free. The team has helped us with pricing and usage options, set up a tailored session booking process, and makes sure everything runs smoothly from start to finish. And honestly, the talent is always top notch!” Looking for voice talent that makes your brand shine? You’ll find it here. Follow us to discover how easy it is to bring your next project to life.

If your AI voice model still sounds flat, it’s probably not the tech. It’s your data. The best-sounding models aren’t trained on scraps. They’re trained on real human speech. Thousands of hours of diverse, emotionally rich voiceover, ethically sourced with clear consent and fair compensation. Because respecting talent doesn’t just sound better. It IS better. Want to build more human AI? Start by respecting the humans behind it. Follow for more.

Missed the launch? No worries, you can still catch it on demand! We’re talking full consent (no fine print), real support for talent, and clean, consistent voice data clients can actually use. It’s quality built from the ground up. And yeah, we’re pretty proud of it. Watch the full Voice Data Launch Event on-demand, link is below. voices.com/landing/voice-dat…