Vibe-Code a Voice AI Agent in 3 Hours — Live Workshop with Claude Code - Open Registration

Join us for an exciting hands-on workshop where you won't just learn how to create a voice agent—you'll shape one into a tailor-made product with the help of Claude Code! 🚀

By the end of the session, you’ll have your own custom voice AI agent up and running on LiveKit Cloud, ready to face 50 synthetic callers and refine with AI debugging.

🍕 What's on the Menu?

• Enjoy delicious pizza and refreshing drinks throughout the workshop! 🍻

How It Works

We kick things off with a pre-built LiveKit voice agent on a robust tech stack: Groq + Deepgram + ElevenLabs, boasting sub-500ms latency and Noveum tracing pre-configured. You’ll select, clone, and reshape this agent into your own creation, whether it be a pizza bot, healthcare triage assistant, or customer support agent for your SaaS product. Your imagination is the limit!

The Voice Stack (Ready to Roll!)

• LiveKit Cloud — Your real-time voice infrastructure with free deployment!
• Deepgram Nova-2 — Streaming speech-to-text system with approximately 150ms response time.
• Groq gpt-oss-120b — Sub-200ms LLM inference.
• ElevenLabs Turbo v2.5 — Crisp, low-latency TTS.
• Noveum — Offering tracing, synthetic users, and AI debugging.

Your goal: keep latency under 500ms for a fluent experience!

Before the event, please clone this repository and follow the provided instructions: GitHub Workshop Starter

Agenda (3 Hours)

• 4:30 – 4:45 — Welcome, introductions, and why voice AI can fail in production (pizza is ready!) 🍕
• 4:45 – 5:15 — 🛠 FORK & DEPLOY — Clone the starter and push it to LiveKit Cloud.
• 5:15 – 5:45 — 🎨 CUSTOMIZE WITH CLAUDE — One quick prompt can transform the agent into your unique product.
• 5:45 – 6:30 — 🎭 STRESS-TEST — Using Claude's assistance to conduct 10-call batches and create personas + scenarios.
• 6:30 – 7:00 — 🔍 DEBUG — Traces will reveal where ASR, LLM, or TTS may have faltered.
• 7:00 – 7:20 — 🤖 FIX — Claude applies AI-suggested corrections, redeploys, and re-runs the test batch.
• 7:20 – 7:30 — Demos, Q&A, and networking with more pizza! 🍻

Pre-Workshop Checklist (25 Minutes)

To maximize our session, we recommend completing these tasks beforehand:

• Install Claude Code or Cursor — npm install -g @anthropic-ai/claude-code or download Cursor.
• Ensure all system dependencies are installed: Python 3.10+, Node 20+, git, and the LiveKit CLI.
• Set up free accounts and API keys (no cost involved) for Anthropic, LiveKit Cloud, Deepgram, Groq, ElevenLabs, and Noveum.
• Confirm you forked the workshop starter—repo link sent 24 hours before the event.
• Bring a creative voice agent idea! The most memorable workshops included a Pokémon trivia bot and a yoga instructor.

What's Included?

• Fresh, hot pizza starting at 4:30 PM with more delivered mid-workshop.
• A variety of drinks, including sparkling water and soda.
• Snacks for the duration of the event.
• Every workspace will have power strips available.
• Strong WiFi, plus a backup hotspot.

Who Should Attend?

This event is perfect for founders, engineers, PMs, and designers looking to build or enhance voice AI products—think customer support bots or voice assistants. No advanced engineering skills necessary; if you know how to prompt Claude Code, you're all set!

Hosted By

Shashank Agarwal — Creator of Noveum.ai, an AI-native testing & observability platform trusted by teams managing thousands of real calls daily.

Special Guests TBA — Expect speakers from LiveKit, ElevenLabs, and/or Deepgram.

RSVP early, as we are limited to just 50 builders! Don’t miss out on this incredible opportunity to learn and network.

Location: Frontier Tower, 10th Floor Annex, 995 Market Street, San Francisco.

Apply for founding citizenship in this innovative community: Frontier Tower Application

For more information, check out Frontier Tower's Website or visit Frontier Tower Notion.