Our 2025 ambition: enable builders everywhere to create voice‑AI agents that easily pass the Turing Test - where every conversation feels real and emotionally alive.

We've been working with Yoto to bring more stories to more kids. Their players are in 200+ countries, but most stories support only three languages. Together, we’re changing this. Launching now: Polish, Romanian, German, Portuguese and Hindi. One day: every language.

I’ve rarely felt a conversation land this close to heart. Tim is a former NFL player, bestselling author, and celebrated broadcaster. Since his ALS diagnosis, he’s continued doing amazing work using Voice AI. We just recorded a podcast together and had an amazing discussion.

June was intense: we shipped a new foundational model, rolled out major upgrades to our Conversational AI platform, and more! We deployed Eleven v3 (alpha) - our most expressive Text to Speech model. It adds full control, 70+ languages and dialogue support. It's an alpha version, but with the right prompting, the results can get breathtaking! Conversational AI platform - build your own voice agent - now supports: - MCP (including streamable HTTP) - allowing AI agents to connect easily to services such as Salesforce, HubSpot, Gmail, and more - Agents & Tools are now easier to create - Widget supports text interaction in addition to audio, giving users greater flexibility and easier input - Improved design across the platform Audio Creative platform - create anything with audio - now comes with: - Voice Design v3 - generate any voice with just a prompt. Control pitch, accent, gender, and emotion; no samples needed - Mobile version for both iOS and Android - a streamlined mini version of the platform Showcase: - @11dotai assistant - a voice-agent prototype that demonstrates what you can build with the Conversational AI platform. It speaks, acts, and integrates with products like Slack, Linear, Perplexity And in addition to enterprises & creators, we’re proud to now work with 2M+ developers, that are also building the future of voice.

goose PR #5, built-in voice dictation using @elevenlabsio and whisper-1:

MCP + Conversational AI → creating integrations for AI agents is now even easier.

Thanks for the shoutout, legend!

Voice AI is notoriously tough to showcase, so even happier to see how others are trying to push Eleven v3’s limits. A few fun ones from the last days: from a tough passage to a beautiful pairing with video models. ↓

Launching on the noisiest day just means we raise the bar. We’re upping the ante: shipping > drama. A few Eleven v3 examples from yesterday ↓

Introducing Eleven v3 (alpha) - our most expressive Text to Speech model yet. - Full emotion control - Multi-speaker dialogue - 70+ languages It’s a research preview: takes a bit of prompting, but when it’s right, the output is breathtaking!