Google Colab Secretsfor Hugging Face and OpenRouter

0:00 Introduction and Colab basics 2:29 What are tokens and why use them 4:54 Creating an OpenRouter token 6:59 Using secrets across notebooks 9:43 Example: Hugging Face token and automation 13:49 Why tokens are safer than passwords Learn how to securely manage API tokens in Google Colab, including creating, storing, and using tokens for services like OpenRouter and Hugging Face. This video demonstrates the process step by step, from generating a disposable token to integrating it into your notebooks. You'll see how to use Colab's Secrets feature to store tokens per user, not per notebook, so you can share notebooks without exposing sensitive keys. The video also covers practical examples, such as connecting to OpenRouter and Hugging Face, and explains why tokens are safer than using your username and password. By the end, you'll understand how to set up tokens with limited scope, expiration, and budget, and how to use them in your own AI and data science projects. This is ideal for students and developers who want to work with external APIs in a secure and efficient way. Key takeaways: Use Colab's Secrets feature to store API tokens per user, not per notebook, so you can share notebooks without exposing keys. Create disposable tokens with short expiration times and credit limits to minimize risk if they are accidentally shared. Tokens allow you to grant limited access, such as read-only or write-only permissions, to specific services or organizations. You can test whether a token works by running a simple cell that checks for the secret and makes an API call. Tokens are safer than using your username and password because they can be scoped, budgeted, and revoked. You can use tokens to automate tasks, like checking for updates in a Hugging Face organization and summarizing changes. Key terms: API token: A limited password created by a service that grants access to specific features or data without exposing your main credentials. Colab Secrets: A feature in Google Colab that securely stores sensitive values like API keys, accessible by name in notebooks. OpenRouter: A service that provides access to multiple AI models through a single API, allowing you to create tokens with expiration and credit limits. Hugging Face: A platform for machine learning models and datasets, often used with tokens to access private repositories or organizations. Runtime: In Colab, the execution environment for your notebook, which can be configured with different hardware like CPU, GPU, or TPU. #GoogleColab #APITokens #OpenRouter Table of Contents: 00:41 - Change runtime GPU 01:28 - Table of contents 01:54 - Colab Data Inspector 02:20 - Cool and easy datasets from Kaggle 03:37 - Colab secrets and tokens 04:56 - Generate an API/token Openrouter 06:06 - How to insert that token into Colab 09:24 - Adding a HuggingFace token 11:17 - Agentic coding in action