AI Detector CLI for Claude, Codex, Hermes Agent, OpenClaw

An AI detector CLI is useful when the check belongs inside the workflow, not in another browser tab. You may be reviewing a folder of screenshots, checking a draft from Terminal, or letting Claude, Codex, Hermes Agent, OpenClaw, or Cursor score text before another rewrite.

Slop or Not now ships two Mac surfaces for that job: the slop CLI and the slop mcp server. The CLI is for direct shell use. The MCP server, built on the Model Context Protocol, lets AI agents call the same local tools themselves. Both ship inside the Mac app bundle, and both require Pro.

The important part is where the scoring happens. AI text detection, AI image detection, readability, and Text Cleanup run locally on your Mac.

What Does the Slop or Not AI Detector CLI Do?

The slop CLI gives Terminal, scripts, and AI coding tools access to the same checks that live in the Mac app. It scores AI text, checks local images, calculates readability, cleans hidden text artifacts, reports install status, and starts the MCP server.

Install Slop or Not for Mac, activate Pro, then put the bundled binary on your PATH:

mkdir -p ~/.local/bin
ln -sf '/Applications/Slop Or Not - AI Fake Detector.app/Contents/MacOS/slop' ~/.local/bin/slop
slop status

The common commands are:

pbpaste | slop text --json
pbpaste | slop readability --json
pbpaste | slop cleanup --json
slop image --json < ~/Desktop/screenshot.png
slop mcp

Use --json when another tool needs stable output. That is the mode Claude, Codex, Hermes Agent, OpenClaw, Cursor, and scripts can read most reliably.

How Do I Use the CLI for AI Text and Image Detection?

Use slop text when you want a local probability verdict for a pasted draft. Use slop image when you want to check a local image file without uploading it to a cloud detector.

Text checks read from stdin:

pbpaste | slop text --json

Image checks read image bytes from stdin:

slop image --json < ~/Desktop/screenshot.png

The image command checks metadata first when supported, including C2PA and IPTC provenance signals used by major generators. If the metadata is missing, Slop or Not falls back to the on-device image model.

That makes the CLI useful for journalists checking source images, moderators reviewing local files, and writers who want to score a draft before they paste it back into an agent.

How Do I Check Readability Score From Terminal?

Use slop readability when you want reading-level feedback without sending the text to another web tool. Readability is separate from AI detection, so it should be used as a writing-control metric rather than an authorship verdict.

pbpaste | slop readability --json

English returns Flesch Reading Ease and Flesch-Kincaid Grade Level. German uses Wiener Sachtextformel. Spanish uses Flesch-Szigriszt. Italian uses GULPEASE. Swedish, Danish, and Norwegian use LIX.

This helps when a rewrite makes the text too formal or too simple. You can ask Claude or Codex to preserve a target reading level, re-run the readability check, and compare the result.

How Do I Clean Hidden Text Artifacts?

Use slop cleanup before a second detector pass when text may contain zero-width characters, homoglyphs, odd spacing, or other mechanical artifacts. Cleanup does not rewrite the argument. It removes characters that should not be part of the prose.

pbpaste | slop cleanup --json

This is useful after copying from a web editor, PDF, or AI tool. It is also useful inside an Agentic Humanizer loop because the agent can clean the text before asking the detector to score it again.

What Does the Slop or Not MCP Server Add?

The Slop or Not MCP server lets Claude, Codex, Hermes Agent, OpenClaw, Cursor, and other MCP clients call Slop or Not as a local stdio server. Instead of copying CLI output back into chat, the agent calls the tools directly.

The server exposes six tools:

MCP tool	What it checks
`slop_status`	App install and Pro state
`detect_text`	AI probability for text
`analyze_readability`	Reading level and readability
`clean_text`	Hidden characters and text artifacts
`detect_image`	Local image files
`score_image`	Raw OmniAID image score

Claude Code can register it like this:

claude mcp add --transport stdio --scope user SlopOrNot -- slop mcp

Codex uses TOML:

[mcp_servers.SlopOrNot]
command = "slop"
args = ["mcp"]

Hermes Agent uses an MCP server entry:

mcp_servers:
  SlopOrNot:
    command: "slop"
    args: ["mcp"]

OpenClaw can register the same local command:

openclaw mcp set slopornot '{"command":"slop","args":["mcp"]}'

The full setup is in the Slop or Not MCP guide.

When Should I Use CLI Instead of MCP?

Use the CLI when you want an explicit command, a shell script, or JSON output you can store. Use MCP when the agent should decide when to call the checker during a writing or review workflow.

CLI is better for:

batch checks from Terminal
local scripts
quick image checks
debugging the app install

MCP is better for:

Claude, Codex, Hermes Agent, or OpenClaw rewrite loops
Agentic Humanizer
Cursor workflows
repeated readability and cleanup checks inside one conversation

Both use the same on-device Slop or Not checks. The difference is who calls the tool: you or the agent.

For Agentic Humanizer, MCP is the smoother path because the agent can score a baseline, clean mechanical artifacts, rewrite, and re-score without copy-pasting JSON. Voice matching still belongs to Agentic Humanizer, not Slop or Not. The optional writing sample steers the rewrite; Slop or Not measures the result locally.

FAQ

Is the Slop or Not CLI free?

The Slop or Not app is free with limited daily checks. The CLI and MCP server require Pro on Mac.

Does the Slop or Not MCP server upload my text or images?

No. Detection, readability, cleanup, and image checks all run locally on your Mac. If an AI agent rewrites text, that agent's own privacy policy applies to the rewrite.

Can I use the Slop or Not CLI without Claude or Codex?

Yes. The CLI works directly from Terminal. Claude, Codex, Hermes Agent, OpenClaw, Cursor, and other tools are optional clients around the same local commands.

Can I use this for Agentic Humanizer?

Yes. Agentic Humanizer uses the Slop or Not MCP server first and falls back to the CLI when needed. If you want rewrites to follow your writing style, read the own-voice AI humanizer guide before you add a sample.

Try It

Install Slop or Not for Mac, activate Pro, then follow the CLI guide or MCP guide. If your goal is rewriting with local feedback, read the Agentic Humanizer guide and the voice-matching guide next.