Skip to content

Instantly share code, notes, and snippets.

View amosgyamfi's full-sized avatar

Amos Gyamfi amosgyamfi

View GitHub Profile
import logging
from dotenv import load_dotenv
from vision_agents.core import User, Agent, cli
from vision_agents.core.agents import AgentLauncher
from vision_agents.plugins import decart, getstream, gemini, elevenlabs, deepgram
logger = logging.getLogger(__name__)

Tech Troubleshooting Vision Agent Instructions

A Tech Troubleshooting Vision Agent Using Baseten and Qwen3-VL


1. Overview

You are a Tech Troubleshooting Vision Agent — an AI-powered hardware repair assistant that helps users diagnose and fix their devices through real-time video analysis and voice guidance.

import logging
from dotenv import load_dotenv
from vision_agents.core import User, Agent, cli
from vision_agents.core.agents import AgentLauncher
from vision_agents.plugins import decart, getstream, openai, elevenlabs, deepgram
logger = logging.getLogger(__name__)
import asyncio
import logging
from uuid import uuid4
from dotenv import load_dotenv
from vision_agents.core.edge.types import User
from vision_agents.core.agents import Agent
from vision_agents.plugins import getstream, elevenlabs, smart_turn, gemini
import asyncio
import logging
from uuid import uuid4
from dotenv import load_dotenv
from vision_agents.core.edge.types import User
from vision_agents.core.agents import Agent
from vision_agents.plugins import getstream, openrouter, elevenlabs, deepgram, smart_turn, moondream
import asyncio
import logging
from uuid import uuid4
from dotenv import load_dotenv
from vision_agents.core.edge.types import User
from vision_agents.core.agents import Agent
from vision_agents.plugins import getstream, deepgram, smart_turn, gemini, cartesia
import asyncio
import logging
from uuid import uuid4
from dotenv import load_dotenv
from vision_agents.core.edge.types import User
from vision_agents.core import agents
from vision_agents.plugins import getstream, ultralytics, gemini, openai

Yoga AI Voice Instructor Guide

  1. Overview

You are a voice yoga instructor — an expert in the art and science of yoga postures (asanas) as defined in the Pocket Yoga Pose Library.

You use YOLO pose analysis to see the practitioner’s exact movements — body alignment, angles, and transitions. Your job is to observe, assess, and guide users through precise, safe, and mindful practice.

Voice and Personality • Speak only in English and with a female voice and a soft American accent — grounding, witty, and slightly snarky when correcting poor form.

"""GitHub MCP Demo - Demonstrate function calling with OpenAI Realtime and GitHub MCP.
This demo shows how OpenAI Realtime can use GitHub MCP tools for real-time function calling
during live conversations. The agent can interact with GitHub repositories, issues, and more
using voice commands through the OpenAI Realtime API.
"""
import asyncio
import logging
import os
import asyncio
import logging
from uuid import uuid4
from dotenv import load_dotenv
from vision_agents.core.edge.types import User
from vision_agents.core.agents import Agent
from vision_agents.plugins import fish, getstream, deepgram, smart_turn, xai