chat

chat

The chat function provides interactive conversation sessions.

Import

from pyai import chat

Basic Usage

# Start chat session
session = chat()

# Send messages
response = session.send("Hello!")
response = session.send("What's your name?")

Quick Start

from pyai import chat

# Create session
session = chat()

# Conversation
print(session.send("Hi, I need help with Python"))
print(session.send("How do I read a file?"))
print(session.send("Can you show an example?"))

# Session maintains context

Parameters

Parameter	Type	Default	Description
system	str	None	System prompt
model	str	None	Model to use
temperature	float	0.7	Response creativity
memory	bool	True	Remember conversation

Examples

Basic Chat

from pyai import chat

session = chat()

while True:
    user_input = input("You: ")
    if user_input.lower() == "quit":
        break
    response = session.send(user_input)
    print(f"AI: {response}")

With System Prompt

session = chat(
    system="You are a helpful Python tutor. Explain concepts simply."
)

response = session.send("What is a decorator?")

With Custom Model

session = chat(
    model="gpt-4",
    temperature=0.5
)

response = session.send("Help me design an API")

Context Manager

from pyai import chat

with chat() as session:
    print(session.send("Hello"))
    print(session.send("Tell me a joke"))
# Session automatically cleaned up

Conversation History

session = chat()
session.send("My name is Alice")
session.send("I work at Acme Corp")

# Later in conversation
response = session.send("What's my name?")
# "Your name is Alice"

# Access history
for msg in session.history:
    print(f"{msg['role']}: {msg['content']}")

Clear and Reset

session = chat()
session.send("Remember: secret code is 1234")

# Clear history
session.clear()

session.send("What's the secret code?")
# Won't remember

Streaming Responses

session = chat()

# Stream response
for chunk in session.stream("Write a long story"):
    print(chunk, end="", flush=True)
print()

Async Usage

import asyncio
from pyai import chat

async def main():
    session = chat()
    response = await session.send_async("Hello!")
    print(response)

asyncio.run(main())

Session Methods

Method	Description
send(message)	Send message and get response
stream(message)	Stream response
clear()	Clear conversation history
save(path)	Save session to file
load(path)	Load session from file

Persistence

session = chat()
session.send("Important context here")

# Save session
session.save("session.json")

# Later, restore session
restored = chat.load("session.json")
restored.send("Do you remember?")

See Also

• ask - Single question answering
• Agent - Full agent framework
• Voice-Module - Voice chat