Package: ai-dialog

All packages Open in Cursor /Users/friedrichstr/ai-mac-studio/agents/communication/ai-dialog

Status

Description: Real-time AI voice dialog using the **Google Gemini Live API**. Talk to Gemini with your microphone; optionally share your camera or screen.

Port: —

Venv: venv present /Users/friedrichstr/ai-mac-studio/agents/communication/ai-dialog/.venv/bin/python

Git: not a git repo

File tree

ai-dialog/
  ai_dialog/
    __init__.py
    __main__.py
    config.py
    core.py
    interpreter.py
  docs/
    reports/
    _quarto.yml
    api.qmd
    architecture.qmd
    cli.qmd
    commands.qmd
    contributing.qmd
    custom.scss
    frontend.qmd
    index.qmd
    intent-recognition.qmd
    konzeption.md
    phases.qmd
  dokumente/
    sprachgesteuerte-textueberarbeitung/
  frontend/
    app.js
    index.html
    styles.css
  demo_dialog.py
  demo_interpret.py
  PROJECT.yaml
  pyproject.toml
  README.md
  server.py

Project Goaldraft

[DRAFT — edit this] Real-time AI voice dialog using Google Gemini Live API Inferred stage: development. Define the specific deliverable, success state, and timeline.

Success criteria:

[DRAFT] Define 3–5 checkable criteria for project completion

Code	`/Users/friedrichstr/ai-mac-studio/agents/communication/ai-dialog`
Docs	`/Users/friedrichstr/ai-mac-studio/agents/communication/ai-dialog/docs`

Path & Service Checks

Check	Path	Status	Detail
code	`/Users/friedrichstr/ai-mac-studio/agents/communication/ai-di`	✓ ok	13 items, last_modified=2026-08-24
docs	`/Users/friedrichstr/ai-mac-studio/agents/communication/ai-di`	✓ ok	20 items, last_modified=2026-08-14

Documentation

README

ai-dialog

Real-time AI voice dialog using the Google Gemini Live API. Talk to Gemini with your microphone; optionally share your camera or screen.

Installation

pip install -e .
# or install dependencies manually:
pip install google-genai opencv-python pyaudio pillow mss

Requirements

Requirement	Notes
Python ≥ 3.10	`asyncio.TaskGroup` required
`GEMINI_API_KEY` env var	Get a key at aistudio.google.com
Microphone + speakers	For bidirectional audio
Webcam	Optional, for `--mode camera`

Usage

# Audio + camera (default)
GEMINI_API_KEY=<key> ai-dialog

# Audio only
GEMINI_API_KEY=<key> ai-dialog --mode none

# Audio + full screen share
GEMINI_API_KEY=<key> ai-dialog --mode screen

Type a message and press Enter to send text alongside voice. Type q to quit.

Python API

import asyncio
from ai_dialog import AudioLoop

asyncio.run(AudioLoop(video_mode="none").run())

Configuration

Edit ai_dialog/config.py to change:

MODEL — Gemini model name
CONFIG — voice, compression, modalities
SEND_SAMPLE_RATE / RECEIVE_SAMPLE_RATE — audio quality

Architecture

ai_dialog/
├── config.py        # Constants + LiveConnectConfig
├── core.py          # AudioLoop class (all async tasks)
├── interpreter.py   # Intent classification
├── __init__.py      # Public API
└── __main__.py      # CLI entry point

Documentation

Full documentation is in docs/ — build with quarto render docs/.

Recent commits

No git history.