Claude Haiku 5.5 Prompt Generator

Write 5 subject lines for an email about [topic]. Requirements: under 45 characters each, no clickbait, no exclamation marks, plain language. The audience is [describe audience]. Output as a numbered list — nothing else.

Classify the sentiment of each review below as Positive, Negative, or Neutral. Output a JSON array with objects containing "id" and "sentiment". Do not explain. [Paste reviews with IDs]

Assign each support ticket below to one of these categories: Billing, Technical, Account, Feature Request, Other. Output only a JSON array: [{"id": "...", "category": "..."}]. Do not explain. [Paste tickets with IDs]

Summarise each item below in exactly one sentence of under 20 words. Output a numbered list matching the input order. [Paste paragraphs or articles]

Extract up to 5 keyword tags from each text below. Return a JSON array: [{"id": "...", "tags": [...]}]. Tags must be lowercase, no punctuation. [Paste texts with IDs]

Write a product description for: [product name]. Key features: [list 3–5 features] Target buyer: [describe buyer] Length: 60–80 words. Tone: clear, confident, no hype. End with one sentence about delivery or guarantee if relevant.

Answer each FAQ below for [company / product name]. Keep each answer under 60 words. Plain language, no jargon. If a question is outside your knowledge, write "Contact our support team at [email]." [List FAQs]

Rewrite the text below to match our brand voice: professional but conversational, active voice, second person ("you"), no passive voice, no exclamation marks. Keep all facts unchanged. Return only the rewritten text. [Paste text]

Convert the data below from [current format, e.g. unstructured list] to [target format, e.g. CSV / JSON / Markdown table]. Column names: [list them]. If a field is missing, use null. Return only the formatted output. [Paste data]

Translate the following from [source language] to [target language]. Keep the original formatting and line breaks. Do not add notes or commentary — return only the translated text. [Paste text]

Build a meeting agenda for a [duration]-minute [meeting type] with these attendees: [list roles]. Topics to cover: [list topics]. Format: numbered agenda items with estimated minutes each, one prep note per item. Allocate time so it totals [duration] minutes exactly.

Generate an SEO-friendly URL slug for each title below. Rules: lowercase, hyphens not underscores, remove stop words, max 7 words. Output a JSON array: [{"title": "...", "slug": "..."}]. [Paste titles]

Draft a reply to each customer message below. Each reply: under 80 words, professional tone, resolve the issue if possible, end with a clear next step. Return numbered drafts matching the input order. [Paste customer messages]

For each product below, generate: a 160-character meta description (include the product name and its main benefit), an alt text for the hero image (under 125 characters, descriptive), and 3 structured data tags. Output as JSON. [Paste product names and short descriptions]

Review each post below. Flag it as Safe, Review, or Remove based on these rules: - Remove: hate speech, illegal content, explicit threats - Review: potentially misleading claims, borderline content - Safe: everything else Output JSON: [{"id": "...", "verdict": "...", "reason": "one sentence"}]. [Paste posts with IDs]

Write a [Python / TypeScript / JavaScript] function that routes incoming [requests / messages / events] to the correct handler based on these rules: [Describe routing rules] Handlers available: [list function names]. The function must return the handler name as a string. Include a default fallback.

Write 10 [Twitter / LinkedIn / Instagram] posts about [topic or campaign]. Each post: [character limit] characters max, hook in the first line, no hashtags unless specified, ends with a call to action. Vary the angle — don't repeat the same point. Output as a numbered list.

Extract the following fields from each document below and return a JSON array. Fields: [list field names and types]. If a field is not present in a document, use null. Do not infer values — only extract what is explicitly stated. [Paste documents]

Build a feature comparison matrix for [product] vs [competitor 1], [competitor 2], [competitor 3]. Features to compare: [list features] Format as a Markdown table. Use ✓ for yes, ✗ for no, and a short note (under 8 words) for partial support. Base the competitor data only on what I provide below. [Paste competitor data]

Rewrite this policy section in plain language for a general audience. Target reading level: US grade 8. Keep every legal requirement — do not remove clauses or change their meaning. Add a one-sentence plain-English summary at the top of each section. Flag any clause you were uncertain how to simplify. [Paste policy text]