Workers AI, Realtime - Pipecat PhoneLLM Mini now available on Workers AI

<p><a href="https://developers.cloudflare.com/workers-ai/models/phonellm-mini-alpha-2/"><code>@cf/daily-co/phonellm-mini-alpha-2</code></a>, a model from the <a href="https://www.pipecat.ai/" target="_blank" rel="noopener">Pipecat<span class="external-link"> ↗︎</span></a> team at Daily, is now available on <a href="https://developers.cloudflare.com/workers-ai/">Workers AI</a>. PhoneLLM is built for voice agents and other realtime AI applications, where a model has to reply quickly and call the right tool mid-conversation.</p> <p>In a voice agent, the language model sits between speech-to-text and text-to-speech, so every second it takes adds to the caller's wait. Developers usually choose between a fast model without reasoning mode, which can miss instructions or tool calls, and a slower model in reasoning mode, which handles both well. Pipecat post-trained NVIDIA Nemotron 3 Nano 30B on phone and customer support scenarios so PhoneLLM follows instructions and calls tools better without reasoning mode.</p> <p>You can use PhoneLLM through the <a href="https://developers.cloudflare.com/workers-ai/configuration/bindings/">Workers AI binding</a> (<code>env.AI.run()</code>) or the <a href="https://developers.cloudflare.com/workers-ai/get-started/rest-api/">REST API</a> at <code>/ai/run</code>. Pipecat recommends <code>temperature: 0</code>. Also set <code>chat_template_kwargs: { enable_thinking: false }</code>.</p> <div tabindex="-1" class="heading-wrapper level-h4"><h4 id="run-a-voice-agent-with-pipecat-on-workers">Run a voice agent with Pipecat on Workers</h4><a class="anchor-link" href="#run-a-voice-agent-with-pipecat-on-workers"><span aria-hidden="true" class="anchor-icon"><svg width="16" height="16" viewBox="0 0 24 24"><path fill="currentcolor" d="m12.11 15.39-3.88 3.88a2.52 2.52 0 0 1-3.5 0 2.47 2.47 0 0 1 0-3.5l3.88-3.88a1 1 0 0 0-1.42-1.42l-3.88 3.89a4.48 4.48 0 0 0 6.33 6.33l3.89-3.88a1 1 0 1 0-1.42-1.42Zm8.58-12.08a4.49 4.49 0 0 0-6.33 0l-3.89 3.88a1 1 0 0 0 1.42 1.42l3.88-3.88a2.52 2.52 0 0 1 3.5 0 2.47 2.47 0 0 1 0 3.5l-3.88 3.88a1 1 0 1 0 1.42 1.42l3.88-3.89a4.49 4.49 0 0 0 0-6.33ZM8.83 15.17a1 1 0 0 0 1.1.22 1 1 0 0 0 .32-.22l4.92-4.92a1 1 0 0 0-1.42-1.42l-4.92 4.92a1 1 0 0 0 0 1.42Z"></path></svg></span></a></div> <p><a href="https://docs.pipecat.ai/" target="_blank" rel="noopener">Pipecat<span class="external-link"> ↗︎</span></a> is an open-source Python framework for building voice agents. Now that <a href="https://developers.cloudflare.com/workers/languages/python/">Python Workers</a> are generally available, you can run a Pipecat voice agent on Cloudflare.</p> <p>Our <a href="https://github.com/cloudflare/realtime-examples/tree/main/pipecat-voice-agent" target="_blank" rel="noopener">Pipecat voice agent example<span class="external-link"> ↗︎</span></a> runs in a <a href="https://developers.cloudflare.com/durable-objects/">Durable Object</a> and uses WebRTC as the voice transport through the <a href="https://developers.cloudflare.com/realtime/sfu/">Realtime SFU</a>. It uses <a href="https://developers.cloudflare.com/workers-ai/models/deepseek-v4-flash-0731/">DeepSeek V4 Flash</a> by default. To deploy it with PhoneLLM, give this prompt to your coding agent:</p> <button type="button" data-nb-button class="group inline-flex w-max shrink-0 items-center justify-center rounded-full font-medium whitespace-nowrap no-underline shadow-xs transition-colors select-none focus-visible:outline-2 focus-visible:outline-ring focus-visible:outline-offset-2 disabled:cursor-not-allowed disabled:opacity-50 bg-card text-foreground ring ring-border hover:bg-accent hover:ring-border-strong h-10 gap-2 px-4 text-sm copy-prompt-btn group/copy cursor-copy" data-prompt="Deploy the Pipecat voice agent example at https://github.com/cloudflare/realtime-examples/tree/main/pipecat-voice-agent with @cf/daily-co/phonellm-mini-alpha-2 as the LLM and thinking disabled. Use the Cloudflare CLI, cf (https://developers.cloudflare.com/cf/), to create Cloudflare resources." data-tooltip="Deploy the Pipecat voice agent with PhoneLLM"><span class="inline-flex items-center gap-2.5 text-base font-medium"><span class="relative flex h-5 shrink-0 items-center justify-start gap-[3px]" aria-hidden="true"><img src="https://developers.cloudflare.com/icons/agents/claude/light.svg" alt="" class="block h-5 w-5 dark:hidden" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/claude/dark.svg" alt="" class="hidden h-5 w-5 dark:block" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/codex/light.svg" alt="" class="block h-5 w-5 dark:hidden" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/codex/dark.svg" alt="" class="hidden h-5 w-5 dark:block" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/cursor/light.svg" alt="" class="block h-5 w-5 dark:hidden" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/cursor/dark.svg" alt="" class="hidden h-5 w-5 dark:block" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/opencode/light.svg" alt="" class="block h-5 w-5 dark:hidden" loading="lazy" decoding="async"><img src="https://developers.cloudflare.com/icons/agents/opencode/dark.svg" alt="" class="hidden h-5 w-5 dark:block" loading="lazy" decoding="async"></span><span class="group-data-[copied=true]/copy:hidden">Copy prompt</span><span class="hidden group-data-[copied=true]/copy:inline">Prompt copied!</span></span></button><script type="module" src="https://developers.cloudflare.com/home/runner/work/cloudflare-docs/cloudflare-docs/src/components/CopyPromptButton.astro?astro&type=script&index=0&lang.ts"></script> <p>For more information, refer to the <a href="https://developers.cloudflare.com/workers-ai/models/phonellm-mini-alpha-2/">PhoneLLM Mini model page</a>, <a href="https://huggingface.co/pipecat-ai/phonellm-mini-alpha-2" target="_blank" rel="noopener">Pipecat's model card<span class="external-link"> ↗︎</span></a>, and <a href="https://developers.cloudflare.com/workers-ai/platform/pricing/">Workers AI pricing</a>.</p>