voice-call-app

Makes voice calls through desktop apps using computer use and live audio with structured data output.

14|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/intendant-dev/Intendant --skill voice-call-app-intendant-dev
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: voice-call-app
Source: https://github.com/intendant-dev/Intendant/tree/main/skills/voice-call-app
Command: npx skills add https://github.com/intendant-dev/Intendant --skill voice-call-app-intendant-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Automating real phone or VoIP calls that require navigating a desktop app's UI and conducting a live AI voice conversation is difficult to coordinate. This Skill orchestrates the full flow: clicking through apps like Element, FaceTime, or WhatsApp to place the call, then running a live audio session that collects typed, structured data from the conversation. ## Core Features & Use Cases - UI-driven call placement: Uses computer use actions to navigate the target app, find the contact, and click the call button, with app-specific guidance such as choosing Legacy call in Element. - Live audio conversation with structured output: The spawn_live_audio session follows a playbook and returns a LiveAudioResult with response_data validated against a mandatory response schema. - Typed response schemas: Define string, integer, boolean, and array fields with required flags, constraints, and tainted markers for user-provided content. - Use Case: Call a restaurant through WhatsApp to confirm a reservation, then immediately write the guest name, party size, confirmed time, and notes to an output file from the structured response. ## Quick Start Ask the agent to call a contact through Element or WhatsApp and collect specific details such as a reservation confirmation, providing the fields you want returned as structured data.

Frequently Asked Questions about voice-call-app

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

FAQPage Schema
How do I make an AI voice call through WhatsApp or Element?▼

Use computer use actions to open the app, navigate to the contact, and click the call button, then call spawn_live_audio with a playbook and response schema in the same turn. The live audio session conducts the conversation and returns structured data.

How do I get structured data from a live voice call?▼

Define a response_schema with typed fields (string, integer, boolean, array) and mark required fields as true. The voice model calls submit_response with the collected data, and the result is returned in LiveAudioResult response_data.

Does this work on Linux or only macOS?▼

It works on macOS and Linux with a display and a supported virtual-audio bridge. macOS prefers Vortex Audio with BlackHole as fallback and requires a GUI session with TCC mic permission; Linux uses PulseAudio null sinks.

Why does the Element call not ring the other person?▼

When clicking the phone icon in Element, a dropdown offers Element Call or Legacy call. Always choose Legacy call, because Element Call is a conference requiring the other party to join manually, while Legacy call rings their device directly.

What happens if the voice call times out or fails?▼

spawn_live_audio returns a status of TimedOut, Disconnected, SchemaError, or Failed depending on the cause. The default timeout is 300 seconds and can be adjusted with the timeout_secs parameter.