notslop

Pull a multi-source social digest (Reddit, Hacker News, blogs, X) on a topic and summarize what people are actually saying right now. Use when the user wants the current conversation around a specific subject.

Use when the user wants to find recent posts semantically similar to a piece of content (URL or text). Embeds the input + ranks candidates across Reddit / HN / blogs / X by cosine similarity. Use for repurpose workflows or "what else is being said about this" research.

Track mentions of a topic across Reddit, Hacker News, blogs, and X over a multi-day window. Use when the user wants to know whether something is getting picked up, how mentions are trending, or to monitor a launch/release over time.

Use when the user has an existing post (tweet, blog, Reddit thread) and wants variants for other platforms, with fresh related context layered in.

Surface what is blowing up right now in a niche across Reddit, Hacker News, blogs, and X. Short-window (6h default) scan reranked by ZeroEntropy. Use when the user wants the current pulse of a space, not a curated read on one topic.

Surface the influential voices (authors, accounts, subreddits) on a topic across Reddit, Hacker News, blogs, and X. Use when the user wants to know who is shaping the conversation, not what is being said.

Use when the user wants 5-10 candidate H1 titles + meta descriptions for a blog post, SEO-optimized for a focus keyword. Faster than `notslop-write-blog-post`; produces titles only, no body.

Use when the user wants to write a long-form blog post (1500-3000 words) on a topic. Pulls a week of social signal, clusters it into themes, identifies undercovered angles, then produces an SEO-aware article with per-section word counts, E-E-A-T citations of real posts, and citation-ready structure for AI Overviews.

Use when the user wants to write a cold DM to someone on Twitter/X or LinkedIn — pulls the recipient's recent posts via notslop and grounds the DM in something specific they said. Outputs a message that respects the recipient's time and references a real thing, not generic flattery.

Use when the user wants a LinkedIn post. Pulls 7-day digest, writes a 1000-1300 char post with line breaks, one hook line, 2-3 data points, takeaway.

Use when the user wants to write a ProductHunt launch — tagline, first comment, and feature highlights. Pulls recent successful PH launches in the same category to ground positioning and ship-quality copy.

Use when the user wants to write or improve the top section of a README for an open-source project — the part above the fold that convinces a developer to keep reading. Pulls competitor READMEs + recent traction signals to ground positioning.

Use when the user wants to write or rewrite their Twitter/X bio (160 char hard cap) or LinkedIn headline. Pulls what's working in their niche to ground positioning.

Use when the user wants an X long-form Article (the "Notes" / Articles format, up to 25k chars). Structured like a blog post with H1/H2 headers, intro, sections, conclusion.

Use when the user wants to draft an X thread (multi-tweet sequence). Pulls digest + clusters + writes 5-8 tweets, each grounded in a specific data point.

Use when the user wants to draft a single tweet on a topic. Pulls fresh signal via notslop, then writes a tweet grounded in 1-2 specific data points. 280-char hard cap.

Use when the user wants to write a new Reddit post (title + body) for a specific subreddit. Adapts tone to the sub culture, grounds in current activity, avoids self-promotion patterns.

Use when the user wants to write a reply to an existing Reddit thread. Reads the thread + sub culture, drafts a comment grounded in current context.

Use when the user wants to write a Show HN launch post for a new project / tool / library / SaaS. Pulls the room temperature of recent successful Show HN posts in the same niche to ground tone + format. Produces a title (<80 chars) and a body that includes problem, what-it-does, stack, demo URL, repo URL.