Workers AI: Image Generation, Text-to-Speech, and the Future of AI Apps

Cloudflare expands Workers AI with Leonardo.Ai and Deepgram models, enabling developers to build low-latency image generation and real-time voice agents.

The world of AI development is rapidly evolving, and Cloudflare is right there in the thick of it. With the expansion of Workers AI to include models from Leonardo.Ai and Deepgram, developers now have powerful tools to build cutting-edge applications.

Image Generation Gets a Boost

Cloudflare's Workers AI is now partnering with Leonardo.Ai, bringing their image generation models to the platform. This is a big deal for anyone looking to create visually stunning applications. Leonardo.Ai's models, like Phoenix and Lucid Origin, are known for their quality and speed. Phoenix excels at text rendering and prompt coherence, while Lucid Origin is fantastic for generating photorealistic images. The speed-to-performance ratio makes them a perfect fit for Workers AI.

According to Peter Runham, CTO of Leonardo.Ai, this collaboration unlocks creative possibilities, from character images for gaming to personalized website visuals.

Voice AI Enters the Chat

But it's not just about visuals. Deepgram's voice AI models are also joining the Workers AI family. Voice is a game-changing interface, offering a higher bandwidth than text and conveying nuances like pacing and intonation. Deepgram's speech-to-text and text-to-speech models are incredibly fast, making them ideal for building low-latency voice agents.

Adam Sypniewski, CTO of Deepgram, emphasizes that hosting their voice models on Cloudflare's Workers AI allows developers to create real-time, expressive voice agents with ultra-low latency. Cloudflare's global network brings AI compute closer to users, enabling lightning-fast conversational AI experiences.

The @cf/deepgram/nova-3 model transcribes audio with high accuracy, while @cf/deepgram/aura-1 generates context-aware speech with natural pacing. Plus, with WebSocket support, developers can keep a live connection to the inference server for bi-directional input and output.

Why This Matters

Cloudflare isn't just offering standalone inference services. They're providing a suite of developer products that allow you to build entire applications around AI. Think Workers for application logic, R2 for storage, and Images for media serving. For voice agents, there's WebRTC and WebSocket support, speech-to-text, text-to-speech, and an orchestration layer via Cloudflare Realtime. Cloudflare is betting big on AI, and they're giving developers the tools they need to succeed.

Looking Ahead

The integration of Leonardo.Ai and Deepgram models into Cloudflare's Workers AI marks a pivotal moment for AI application development. It's about making AI more accessible, faster, and more versatile. This is more than just a set of tools; it's a platform for innovation.

Ready to dive in? Check out the developer docs to get started with these new partner models on Workers AI. Who knows, you might just build the next big thing.