voice-ai

Orchestrates phone calls and real-time voice conversations using Twilio and OpenAI Realtime API.

Updated Mar 23, 2026
One-click install
npx skills add https://github.com/sakamotomomotaro0809-netizen/tateyomi --skill voice-ai-sakamotomomotaro0809-netizen
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: voice-ai
Source: https://github.com/sakamotomomotaro0809-netizen/tateyomi/tree/main/taisun_agent/.claude/skills/voice-ai
Command: npx skills add https://github.com/sakamotomomotaro0809-netizen/tateyomi --skill voice-ai-sakamotomomotaro0809-netizen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Automating phone-based communication—outbound calls, inbound call handling, SMS, and voice broadcasts—normally requires building and maintaining complex telephony infrastructure. This Skill connects Twilio's phone network with the OpenAI Realtime API so an AI agent can place, receive, and conduct live voice conversations without manual telephony wiring. ## Core Features & Use Cases - Outbound & Inbound Call Automation: Place calls with a defined prompt and voice, handle incoming calls as an IVR replacement, and end or monitor active calls. - Realtime AI Conversations: Bridge Twilio Media Streams (g711_ulaw) with the OpenAI Realtime API over WebSocket for live two-way voice dialogue with configurable system prompts, voices, and function tools. - Messaging & Broadcast: Send SMS, voice messages, and one-to-many voice broadcasts for notifications and campaigns. - Use Case: A sales team configures an AI SDR follow-up call: the agent dials a lead, speaks Japanese with the "shimmer" voice, answers product questions via a realtime session, and logs the conversation for review. ## Quick Start Ask the agent to place an outbound call to a phone number with a Japanese greeting prompt using the voice-ai skill, after setting the Twilio and OpenAI environment variables.

Frequently Asked Questions about voice-ai

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make an automated outbound phone call with AI?▼

Use the voice_make_call tool with the recipient's phone number, a prompt describing the conversation goal, and a voice configuration such as voice and language. The call runs through Twilio and connects to an OpenAI Realtime session for live dialogue.

How do I connect Twilio to the OpenAI Realtime API for voice calls?▼

Twilio answers the call and opens a Media Stream over WebSocket carrying g711_ulaw audio. A bridge relays that audio to the OpenAI Realtime API WebSocket, and the AI's audio responses stream back through Twilio to the caller.

What environment variables are required for Twilio voice AI integration?▼

You need TWILIO_ACCOUNT_SID, TWILIO_AUTH_TOKEN, TWILIO_PHONE_NUMBER, OPENAI_REALTIME_API_KEY, and WEBHOOK_BASE_URL for receiving webhooks. The voice-ai MCP server must also be configured.

Can the AI voice agent handle inbound calls instead of an IVR?▼

Yes. Incoming calls to the Twilio number are routed via TwiML WebSocket into a realtime session, where the AI answers questions using a configured system prompt and optional function tools such as order status lookups.

Does the voice AI support languages other than English?▼

Yes. The examples configure Japanese conversations by setting the language parameter to "ja" and choosing a voice such as "shimmer", with prompts written in Japanese for customer-facing calls.

What are the limitations of AI phone call automation with Twilio?▼

It depends on external services: a paid Twilio number, OpenAI Realtime API availability, and a publicly reachable webhook URL. Audio is constrained to the g711_ulaw codec, and call quality depends on network latency between the bridges.