A context window is the amount of text an AI model can hold in mind at one time, measured in tokens. Everything in the chat counts: your instructions, pasted resumes, the job brief and the model's own replies. When the total runs past the limit, the earliest material drops out or gets less attention.
AItopic
2 minread
2related terms
In one line
How much text a model can hold in mind at once, counted in tokens. Paste in more than fits and it starts losing the earliest parts, such as the job brief at the top of a long chat.
Think of the context window as the model's working desk. Anything on the desk can shape the answer. Anything that falls off the edge is gone for that conversation. Tokens are chunks of text, roughly a short word or part of a longer word, so a long job description plus ten resumes can fill more space than you expect.
Larger windows let you paste whole interview transcripts or a full requisition history in one go. That does not mean the model weighs every line equally. Material buried in the middle of a very long chat tends to get less attention than the start and the most recent messages.
Most recruiting tools hide the window from you. Your ATS assistant or AI note taker decides what to send to the model behind the scenes, which is why the same question can get a sharper answer in a fresh chat with only the relevant facts.
Why it matters to recruiters
Recruiters lean on AI for long, messy work: comparing a stack of candidates against an intake brief, summarizing a full day of interviews, drafting a search update. If the brief scrolls out of the window, the model starts ranking candidates on general impressions instead of your must haves, and you may not notice.
An example from a recruiting desk
A sourcer pastes a hiring manager's intake notes, then works through 40 profiles in the same chat over an afternoon. By profile 30, the summaries stop mentioning the required security clearance. The fix is simple: start a new chat, paste a short version of the brief at the top, and review the remaining profiles in batches of ten.
Tips that help
Put the most important instructions, such as must have skills, in a short block at the top of every new chat.
Start a fresh chat for each requisition instead of running one endless thread.
Trim pasted documents to the parts that matter, such as the skills section of a resume, before you paste them.
Ask the model to restate the brief before it ranks candidates, so you can see whether it still has it.
Mistakes to avoid
Assuming a bigger window means the model reads every line with equal care.
Mixing two different roles in one chat, so details from one req leak into the other.
Frequently asked questions
What happens when you go over the context window?
The tool either refuses the input or quietly drops the oldest parts of the conversation. Either way the model no longer sees that text, so answers can lose details you gave early on. Starting a new chat with a short summary is the usual fix.
Is a context window the same as memory?
No. The context window covers the current conversation only. Some assistants add a separate memory feature that saves facts across chats, but that is stored outside the window and pulled back in when it seems relevant.
How many resumes can I paste into one AI chat?
It depends on the model and on resume length, so there is no fixed number. In practice, quality drops well before the hard limit. Smaller batches with the brief restated at the top give more consistent results than one huge paste.
