Google Colab Example

It is possible to use ProxAI in Google Colab.

It is recommended to directly follow the instructions from following colabs instead of this documentation page.
🏃‍♂️‍➡️ ProxAI Quick Start Tutorial 🏃‍♂️‍➡️: Good for cold starts
🚀 ProxAI Advanced Usage Tutorial 🚀: Good after spending some time with ProxAI

Installation

Run the following code in a Google Colab cell to install ProxAI:

!pip install proxai

After installation, you need restart the session as instructed in the output of the installation cell:

Get Provider Keys

ProxAI works with various providers like Google Gemini, OpenAI, Anthropic, etc. See our provider integrations page for the full list and setup instructions. Add your API keys to the Google Colab secrets.

Run following code after adding the secrets to the notebook environment:

import os
from google.colab import userdata
 
API_KEY_LIST = [
    'GEMINI_API_KEY',
    'OPENAI_API_KEY',
    'ANTHROPIC_API_KEY',
    'XAI_API_KEY',
    'DEEPSEEK_API_KEY',
    'MISTRAL_API_KEY',
    'CO_API_KEY',
    'DATABRICKS_HOST',
    'DATABRICKS_TOKEN',
    'HUGGINGFACE_API_KEY',
    'PROXDASH_API_KEY',
]
 
for api_key in API_KEY_LIST:
  try:
    os.environ[api_key] = userdata.get(api_key)
  except:
    continue

Note: This is the safest way to store API keys in Google Colab. When you share the Colab notebook with others, these keys are not visible to them.

Start Building

It is ready to use:

import proxai as px
 
provider_models = px.models.list_models(only_largest_models=True)
 
for provider_model in provider_models:
  answer = px.generate_text(
      prompt='Hello model',
      provider_model=provider_model)
  print(f'{provider_model}: {answer}')