Skip to content

Latest commit

 

History

History
77 lines (52 loc) · 1.57 KB

quickstart.md

File metadata and controls

77 lines (52 loc) · 1.57 KB

Quickstart

Get started in 5 minutes.

Install

pip install miiflow-agent

Set your API key:

export OPENAI_API_KEY="sk-..."

First Chat

The basic pattern: create a client, send messages, get responses.

from miiflow_agent import LLMClient
from miiflow_agent.core import Message

client = LLMClient.create("openai", model="gpt-4o-mini")
response = client.chat([Message.user("What is Rust?")])
print(response.message.content)

Streaming

Same interface, get chunks instead of full response:

for chunk in client.stream_chat([Message.user("Explain async/await")]):
    print(chunk.delta, end="", flush=True)

Switch Providers

Change one line, everything else stays the same:

# OpenAI
client = LLMClient.create("openai", model="gpt-4o-mini")

# Claude
client = LLMClient.create("anthropic", model="claude-3-5-sonnet-20241022")

# Groq (fast inference)
client = LLMClient.create("groq", model="llama-3.3-70b-versatile")

# Same interface for all
response = client.chat([Message.user("Hello")])

Async

import asyncio

async def main():
    client = LLMClient.create("openai", model="gpt-4o-mini")

    response = await client.achat([Message.user("Hi")])
    print(response.message.content)

    async for chunk in client.astream_chat([Message.user("Count to 10")]):
        print(chunk.delta, end="", flush=True)

asyncio.run(main())

Next