A drop-in block for Claude's account-level Instructions that blocks 150+ AI slop patterns. Sound human in every Claude conversation, Artifact, and Cowork session.
Go to claude.ai, click your name in the bottom-left, then choose Settings.
Open General and scroll to Instructions for Claude. Paste the slopt block and save.
Every chat (and Cowork session) inherits the rules automatically. No per-conversation paste.
Custom instructions for every model, plus the raw skill.md. Delivered to your inbox automatically.
17% off for the first 250 customers (5 left)
Both. On Claude.ai paste the block into Settings → General → Instructions for Claude — it applies across every chat and Cowork. On the API, pass the slopt block as the system prompt parameter on each request.
Yes. The rules target stylistic patterns, not specific model versions. Sonnet, Opus, Haiku — all the same.
Yes. Append slopt's block to whatever you already have in Instructions for Claude. slopt is purely about voice, so it doesn't conflict with task-specific guidance.
You get paste-ready Claude instructions, the broader custom-instructions pack for ChatGPT, Claude, Gemini, Cursor, Codex, Notion AI, Perplexity and Copilot, plus the same rules as an editable skill.md file.
A humanizer rewrites Claude's output after the slop is already there. slopt sits in the instructions layer first, so Claude is steered away from em-dashes, filler phrases, canned openings, and other patterns before the draft is written.
No prompt can guarantee that Claude will never use a blocked pattern again. You may still see the occasional em-dash, filler phrase, or canned structure, especially after a model update or in a long answer. slopt's job is to push Claude away from those habits at the source, so the pattern shows up far less often and the voice moves in the right direction.
Yes. The Claude block is plain text, and the included skill.md is plain markdown. You can add your own voice rules, remove patterns you like, or layer slopt under an existing brand voice guide.