—pip install →
GantryGraph / Quickstart

Quickstart

From zero to a working agent in five minutes.

Requirements
Python 3.10+ · macOS, Linux, or Windows · an API key from Anthropic, OpenAI, or any LangChain-compatible provider.

1. Install

pip install gantrygraph

# with browser and search extras
pip install 'gantrygraph[browser,search]'
playwright install chromium

2. Set your API key

export ANTHROPIC_API_KEY="sk-ant-..."
# or
export OPENAI_API_KEY="sk-..."

3. Your first agent

Create hello_agent.py:

from gantrygraph import GantryEngine
from langchain_anthropic import ChatAnthropic

agent = GantryEngine(
    llm=ChatAnthropic(model="claude-sonnet-4-6"),
    max_steps=20,
)

result = agent.run("What is the current working directory?")
print(result)

4. Add a browser

from gantrygraph import GantryEngine
from gantrygraph.perception import WebPage
from gantrygraph.actions import BrowserTools
from langchain_anthropic import ChatAnthropic

web = WebPage(url="https://news.ycombinator.com", headless=True)

agent = GantryEngine(
    llm=ChatAnthropic(model="claude-sonnet-4-6"),
    perception=web,
    tools=[BrowserTools(web_page=web)],
    max_steps=20,
)

result = agent.run("Find the top 5 stories and return their titles and links.")
print(result)
That's it
You have a working browser agent. From here, add tools, tighten security, and stream events to your UI.

5. Stream events in real time

import asyncio
from gantrygraph import GantryEngine
from gantrygraph.perception import WebPage
from gantrygraph.actions import BrowserTools
from langchain_anthropic import ChatAnthropic

async def main():
    web = WebPage(url="https://example.com", headless=True)
    agent = GantryEngine(
        llm=ChatAnthropic(model="claude-sonnet-4-6"),
        perception=web,
        tools=[BrowserTools(web_page=web)],
    )
    async for event in agent.astream_events("Find the pricing page"):
        print(event.event_type, event.step, event.data)

asyncio.run(main())

Next steps

← PreviousIntroduction