Positioning, onboarding, changelogs and developer-facing copy — written so a technical audience doesn't bounce off the marketing tone.
Each prompt is reviewed before it goes live. Copy it, replace the bracketed placeholders with your own details, and keep the parts that already match how your team writes.
You are a direct-response copywriter with 15 years of B2B SaaS experience. Product: {PRODUCT} Audience: {ICP} Primary pain: {PAIN} Write landing page copy with: 3 headline options (max 9 words), a 2-sentence subhead, 3 benefit blocks (headline + 20 words), objection-handling FAQ (5 Q&A), and a CTA. Use concrete outcomes and numbers, avoid hype adjectives, 8th-grade reading level.
You are an SEO strategist. Target keyword: {KEYWORD}. Search intent: {INTENT}. Produce a page brief with: recommended H1, URL slug, meta title (<60 chars), meta description (<155 chars), H2/H3 outline covering entity coverage, 8 questions to answer, 5 internal link targets, and a schema.org type recommendation. No fluff, output as markdown.
Act as a staff engineer reviewing a pull request. Here is the diff: {DIFF} Return: (1) blocking issues with file/line and a fix, (2) non-blocking suggestions, (3) missing tests, (4) security or performance risks. Be terse, no praise, code snippets only where they change behavior.
Stack trace: {TRACE} Context: {STACK / FRAMEWORK}. Diagnose the most likely root cause, list 3 alternative causes ranked by probability, tell me exactly what to log or inspect to distinguish them, and give the minimal fix. No speculation presented as fact.
Here is a legacy module: {CODE} Refactor it without changing behavior. Deliver: the refactored code, a list of behavior-preserving changes, any behavior you could not preserve and why, and the characterization tests you would write first. Prefer small named functions over comments.
Real outputHere is a cohort retention table: {TABLE} Act as a growth analyst. Identify the 3 most important patterns, quantify each with the actual numbers, name the most likely cause of the biggest drop, and propose 3 experiments ranked by expected impact and effort. Be skeptical of small-sample cohorts.
Video still — real outputFrom these 20 recent tickets: {TICKETS} Cluster them into at most 8 recurring issues. For each, write a reusable macro (under 120 words) with placeholders, a suggested title, and the trigger phrases a routing rule should match. Output as a table.
