intermediate16 min read· Module 7, Lesson 2
📋Prompt Templates & Patterns
Copy-paste ready templates for summarization, extraction, classification, and more
Prompt Templates & Patterns
Instead of writing prompts from scratch every time, use proven, effective templates. In this lesson, you'll find ready-to-use templates you can copy and adapt.
1. Summarization Template
Summarize the following text in [number] key points.
Rules:
- Each point should be a single clear sentence
- Focus on the most important information
- Ignore secondary details
Text:
"""
[paste text here]
"""
Code example:
async function summarize(text, numPoints = 5) {
const response = await client.messages.create({
model: "claude-sonnet-4-20250514",
max_tokens: 1024,
messages: [{
role: "user",
content: `Summarize the following text in ${numPoints} key points.\n\n"""${text}"""`
}],
});
return response.content[0].text;
}2. Data Extraction Template
Extract the following from the text and return as JSON:
- name
- email
- phone
- company
If a field is not found, use null.
Return valid JSON only, no additional text.
Text:
"""
[paste text here]
"""
Code example:
async function extractInfo(text, fields) {
const response = await client.messages.create({
model: "claude-sonnet-4-20250514",
max_tokens: 1024,
messages: [{
role: "user",
content: `Extract ${fields.join(", ")} from the text as JSON. Use null for missing fields.\n\n"""${text}"""`
}],
});
return JSON.parse(response.content[0].text);
}3. Classification with Few-Shot
async function classifySentiment(text) {
const response = await client.messages.create({
model: "claude-haiku-3-5-20241022",
max_tokens: 50,
messages: [{
role: "user",
content: `Classify sentiment as: positive, negative, or neutral
Examples:
"Great product, highly recommend!" — positive
"Terrible, would never buy again" — negative
"It's okay, nothing special" — neutral
Now classify:
"${text}" —`
}],
});
return response.content[0].text.trim();
}4. Code Review Template
Review the following code for:
1. Logic errors
2. Security issues
3. Performance problems
4. Readability
For each issue: severity, line number, description, suggested fix with code.
5. Translation Template
Translate from [source language] to [target language].
Rules:
- Maintain the same tone and style
- Do not translate proper nouns or technical terms
- Preserve formatting
6. Document Q&A Template
You answer questions based ONLY on the provided document.
Rules:
- Answer only from the document
- If not found, say "This information is not in the document"
- Quote relevant sections when answering
7. Reusable Template Library
const templates = {
summarize: (text, n) => `Summarize in ${n} points:\n"""${text}"""`,
classify: (text, categories) =>
`Classify into: ${categories.join(", ")}. Reply with just the label.\n\n"${text}"`,
extract: (text, fields) =>
`Extract ${fields.join(", ")} as JSON. Use null for missing.\n\n"""${text}"""`,
translate: (text, from, to) =>
`Translate from ${from} to ${to}. Keep proper nouns and code.\n\n"""${text}"""`,
};
// Usage
const prompt = templates.classify(
"The product broke after one day",
["positive", "negative", "neutral"]
);8. Step-by-Step Explanation Template
Explain [topic] step by step.
Rules:
- Assume the reader has no background
- Use simple language
- Include an analogy or real-world example
- End with a summary
Tips for Effective Templates
- Be specific — Define exactly what you want and don't want
- Give examples — Examples are more powerful than instructions
- Specify format — Say how you want output (JSON, bullets, table...)
- Use delimiters — Separate instructions from data with
"""or--- - Test and iterate — Try the template with different data and refine
Summary
Ready-made templates save time and improve result quality. Save your favorite templates and use them as starting points for every new task.
Next: We'll learn how to build AI agent loops.