What Is AI Context?

In plain language

Context is the information an AI can use while working on the current request. It may include:

• your instructions;
• earlier messages;
• uploaded files;
• project code;
• information returned by tools;
• system instructions.

A closely related term is context window: the amount of information a model can process in one prompt. Google defines it in terms of how many Tokens a model can process. (Google Cloud Documentation)

An everyday example

An analogy

A larger desk can hold more material, but that does not mean every document belongs on it. This is an editorial analogy, not a technical definition.

Where you will encounter it

You will see context discussed in long conversations, coding agents, uploaded documents, Skills, tool definitions and model limits. It matters whenever an AI needs to keep track of instructions or work across a large amount of information.

Should you care

• All AI users: worth understanding.
• Codex and agent users: important for long or complex tasks.
• Heavy MCP users: useful when deciding which tool definitions and outputs belong in a session.
• People building reusable Skills: useful when deciding how much guidance should be loaded.

How it differs

Context vs context window: Context is the information available for the current task. The context window is the limit on how much information the model can process at one time.

Context vs memory: Context is not necessarily permanent memory. It describes what is available for the current work.

Context vs token: Tokens are the units used to process content. A context window is commonly measured in tokens.

What to read next

Read What is a token? or compare how context differs between Playwright MCP and the CLI.