Docs managed by Standup StevoEvery content on Liners — product descriptions, taglines, FAQs, and blog posts — is generated by AI with an extreme process of research and quality assurance that's almost impossible for humans to emulate without going on sick leave 50 times a year.
When a product is added to our platform, we generate several pieces of content automatically:
The product description (50-300 words) is generated using two system prompts:
Your brand here — Your competitors haven't found this spot yet. You're welcome.content_system: Sets the overall writing style, tone, and guidelines.content_generation: Provides the specific instructions for generating a product description.The AI receives the product's name, website URL, and any existing information, then produces a factual, well-written description. We make sure the description avoids marketing fluff and focuses on what the product actually does.
A short tagline (under 100 characters) is generated alongside the description. The tagline captures the product's core value proposition in a single line. It never ends with a period — that's one of Quinn's 47 QA checkpoints.
Product FAQs are generated using a two-step process:
faq_generation prompt, a search-enabled AI model researches common questions about the product by searching the web.faq_fallback prompt generates FAQs based on the product description alone.FAQs help users quickly understand key details like pricing, supported countries, integrations, and use cases.
Our structured_output_hints system prompt provides formatting guidelines for AI-generated content. It includes rules for:
We generate blog posts using a similar two-phase approach to comparisons:
blog_research prompt, a search-enabled AI gathers current information about the topic.blog_system (editorial guidelines) and blog_article (structure), the AI produces a full article.Our blog posts follow SEO best practices with proper heading hierarchy, keyword targeting, and internal linking.
We use a set of tunable system prompts to control content generation. Here's a summary of the content-related ones:
| Prompt Key | Purpose |
|---|---|
content_system | Writing style and tone guidelines. |
content_generation | Product description generation instructions. |
structured_output_hints | Formatting rules for taglines, descriptions, content. |
faq_generation | FAQ research instructions. |
faq_fallback | Backup FAQ generation from description only. |
blog_research | Blog post research guidelines. |
blog_system | Blog editorial voice and standards. |
blog_article | Blog post structure and format. |
category_suggestion | Category assignment instructions. |
tag_suggestion | Tag assignment instructions. |
country_suggestion | Country identification instructions. |
Having prompts be tunable means we can self0improve by refining content quality and style over time without changing any code.
All generated content passes through 
QA Quinn's QA process before going live. Quinn checks for:
We generate content on Liners using AI with tunable system prompts. All content is reviewed by Quinn before publication.