Agent demo

A small demo of an LLM agent with tool use: single Python file, OpenRouter API, and a few tools (time, math, book meeting). Built to show the basics of function calling and the agent loop in one place.

What it does

• Chat loop with an assistant (via OpenRouter).
• Tools: get current time, evaluate math expressions, “book” a meeting (name + time).
• Optional TTS on macOS using say (toggle with TTS_ENABLED in agent.py).

Setup

python -m venv .venv
source .venv/bin/activate   # or .venv\Scripts\activate on Windows
pip install -r requirements.txt

Set your OpenRouter API key:

export OPENROUTER_API_KEY=your_key_here

Get a key at openrouter.ai.

Run

python agent.py

Then type messages; Jimmy can use the tools (e.g. “What time is it?”, “What’s 2 + 3 * 4?”, “Book a meeting with Alice tomorrow at 3pm”) and you’ll see tool calls and results in the terminal.

Project layout

• agent.py — agent loop, tool definitions, and tool runner (all in one file for the demo).
• requirements.txt — openai (used against OpenRouter’s OpenAI-compatible API).