Definition
The context window is the amount of text a model can hold in mind at once, the prompt plus retrieved documents plus its own answer. Anything beyond it falls out of view.
On this page
At a glance
- Categories Technical
- Related fields AI Search
- Difficulty Intermediate
How Context Window Works
Think of it as the model's working memory for a single response. It is measured in tokens, and it is finite. When an engine grounds an answer, it has to fit the most relevant passages into that window, which is one reason retrieval picks specific chunks rather than whole sites.
Why Context Window Matters
Because space is limited, engines favor tight, self-contained passages they can slot in efficiently. Long, rambling pages are harder to use than ones broken into clean, liftable sections.
Frequently Asked Questions
-
What happens if content exceeds the context window?
It gets truncated or left out. The model only works with what fits.
-
How does this affect my content?
It rewards clear, self-contained sections that an engine can pull in without dragging along the whole page.