ChatGoogle

ChatGoogle(
    system_prompt=None,
    model=None,
    reasoning=None,
    api_key=None,
    kwargs=None,
)

Chat with a Google Gemini model.

Prerequisites

NoteAuthentication

The simplest way to authenticate is with an API key. Sign up for an account and get an API key, then pass it via api_key or set the GOOGLE_API_KEY environment variable.

For OAuth, service accounts, or Application Default Credentials, pass a google.auth.credentials.Credentials object via kwargs:

import google.auth

credentials, _ = google.auth.default()
chat = ChatGoogle(kwargs={"credentials": credentials})

NotePython requirements

ChatGoogle requires the google-genai package: pip install "chatlas[google]".

Examples

import os
from chatlas import ChatGoogle

chat = ChatGoogle(api_key=os.getenv("GOOGLE_API_KEY"))
chat.chat("What is the capital of France?")

Parameters

Name	Type	Description	Default
system_prompt	Optional[str]	A system prompt to set the behavior of the assistant.	None
model	Optional[str]	The model to use for the chat. The default, None, will pick a reasonable default, and warn you about it. We strongly recommend explicitly choosing a model for all but the most casual use.	None
reasoning	Optional['int | ReasoningEffort | ThinkingConfigDict']	If provided, enables reasoning (a.k.a. “thoughts”) in the model’s responses. This can be an integer number of tokens to use for reasoning (a thinking budget), a string thinking level ("minimal", "low", "medium", or "high"), or a full ThinkingConfigDict to customize the reasoning behavior.	None
api_key	Optional[str]	The API key to use for authentication. You generally should not supply this directly, but instead set the GOOGLE_API_KEY environment variable.	None
kwargs	Optional['ChatClientArgs']	Additional arguments to pass to the genai.Client constructor.	None

Returns

Name	Type	Description
	Chat	A Chat object.

Note

Pasting an API key into a chat constructor (e.g., ChatGoogle(api_key="...")) is the simplest way to get started, and is fine for interactive use, but is problematic for code that may be shared with others.

Instead, consider using environment variables or a configuration file to manage your credentials. One popular way to manage credentials is to use a .env file to store your credentials, and then use the python-dotenv package to load them into your environment.

pip install python-dotenv

# .env
GOOGLE_API_KEY=...

from chatlas import ChatGoogle
from dotenv import load_dotenv

load_dotenv()
chat = ChatGoogle()
chat.console()

Another, more general, solution is to load your environment variables into the shell before starting Python (maybe in a .bashrc, .zshrc, etc. file):

export GOOGLE_API_KEY=...