You can give Claude many types of text—articles, emails, meeting transcripts, database records—and it can help you digest, explain, and answer questions about them. With its 100K or 200K token [context window](🔗) (depending on which model) Claude can analyze tens of thousands of words.

Here are a few basic applications of this capability.

## Evaluating text similarity

We can ask Claude if two pieces of text are roughly the same in meaning.

UserYou are going to be checking whether two sentences are roughly saying the same thing. Here's the first sentence: "{{SENTENCE1}}" Here's the second sentence: "{{SENTENCE2}}" Please begin your answer with "[YES]" if they're roughly saying the same thing or "[NO]" if they're not.

By starting Claude's response ourselves with `[`, we help "reinforce" the prompt instruction to use that format and to start its response with yes or no.

## Answering questions about a text

Here, we give Claude a meeting transcript and a question for it to answer using the transcript.

UserI'm going to give you an example transcript from a meeting and then I'm going to ask you some questions about the transcript. \<transcript> {{TEXT}} \</transcript> Here is the first question: {{QUESTION}}

Citing sources

For an example prompt where we ask Claude to answer a question based on a document _and_ cite sources for its answer, see [Advanced Text Analysis](🔗).