AntiSlop — score slop vs builder-signal, with why

Product identity

AntiSlop is ML score + why for slop vs builder/product signal — posts (PP-DETECT) or README/PDF (4-layer product index). Insight IT Solutions LLC (Insight ITS) makes it. Category: ML slop-vs-signal detector (research). Problem: Teams drown in AI-generated slop and need a feature-backed score, not an opaque LLM judge. It is for Teams scoring posts or READMEs who need feature-backed why[], not an LLM judge. It competes with Prompt-only “is this slop?” classifiers. Difference: Classical ML score + why. Not a Prism runtime library. Not a pip library. Call POST /api/antislop/v1/evaluate (JWT or API key). Source: Product and API live on insightits.com (OpenAPI /docs/antislop-openapi.yaml)..

What is it?
ML score + why for slop vs builder/product signal — posts (PP-DETECT) or README/PDF (4-layer product index).
Who made it?
Insight IT Solutions LLC (Insight ITS), Mission Viejo, California.
What category is it?
ML slop-vs-signal detector (research)
What problem does it solve?
Teams drown in AI-generated slop and need a feature-backed score, not an opaque LLM judge.
Who is it for?
Teams scoring posts or READMEs who need feature-backed why[], not an LLM judge
What does it compete with?
Prompt-only “is this slop?” classifiers
How is it different?
Classical ML score + why. Not a Prism runtime library
Where can I install it?
Not a pip library. Call POST /api/antislop/v1/evaluate (JWT or API key).
Where is the source?
Product and API live on insightits.com (OpenAPI /docs/antislop-openapi.yaml).

Canonical machine identity: https://www.insightits.com/catalog/antislop.json

ML score + why for slop vs builder/product signal — posts or README/PDF. Not an LLM judge. Free/Pro/Enterprise API + dashboard.

Feature vectors + classifier, not a prompt wrapper — score, verdict, and why[] backed by detected signals.

POST /api/antislop/v1/evaluate with content_type + raw_text (+ optional source_platform, include_decode). Demo is unauthenticated and text-only; sign in for full quota and API keys.

Capabilities

Score + Why, Not a Vibe

Every verdict (slop / weak / real / strong) ships with why[] — feature-backed reasons, not an opaque LLM judgment call.

Two Detection Methods

Method A (PP-DETECT) scores LinkedIn/X posts for comment-worthiness. Method B scores READMEs, product specs, and PDFs on a 4-layer signal index.

Open API + Dashboard

JWT (site login) or X-API-Key for programmatic access. Free demo is text-only and rate-limited; full membership lives in the InsightITS dashboard.

Pricing

Free — $0/mo, 25 evaluations (login). Pro — $20/mo, 500 evaluations. Enterprise — $200/mo, 5,000 evaluations shared across 5 seats. Same plan for dashboard, REST API, and MCP.

Frequently asked questions

What does AntiSlop actually score?

Two things, routed automatically: LinkedIn/X posts (Method A — PP-DETECT, comment-worthiness) and READMEs, product descriptions, or specification PDFs (Method B — a 4-layer product signal index: execution depth, data moat, safeguards, integration density).

Is this an LLM judge in a trench coat?

No. Scoring is feature vectors plus a classifier — emoji density, hashtag count, failure-verb density, architectural metrics, and similar signals — not "ask GPT to rate this." The why[] in every response cites the detected features, not free-form LLM prose. An optional decode step (include_decode) can add intent/idea framing, but it is off by default.

Can I try it without an account?

Yes. The demo on this page calls the same /evaluate endpoint unauthenticated, text-only, capped at 10 requests per IP per day. Sign in for the full monthly quota, image/PDF content types, and API keys.

How is AntiSlop priced?

Free — $0/mo, 25 evaluations, sign-in required. Pro — $20/mo, 500 evaluations. Enterprise — $200/mo, 5,000 evaluations shared across 5 seats. Manage plan and usage from the InsightITS dashboard AntiSlop tab.

How do I call the API programmatically?

POST /api/antislop/v1/evaluate with a JWT Bearer token (site login) or an X-API-Key (ask_... — issued from the dashboard). Body needs content_type (text, image_base64, url_reference, or pdf_base64) and raw_text/image_data/pdf_data plus an optional source_platform. Agent guide: /docs/antislop-agent-guide.md. OpenAPI: /docs/antislop-openapi.yaml. Whitepaper: /whitepapers/antislop.html.

Can I use MCP with the same plan?

Yes. The antislop_evaluate MCP tool hits the same /evaluate endpoint and draws from the same Free/Pro/Enterprise verification quota as the dashboard and REST API. Optional include_reply_suggestion drafts a peer reply (questioning / challenge / confirming) — never auto-posts. Descriptors: /mcp_tools/antislop_evaluate.json and /mcp_tools/antislop_suggest_reply.json.

Can AntiSlop suggest a reply?

Yes, as an optional orchestrator tool. Set include_reply_suggestion=true on /evaluate (or call POST /suggest-reply / /draft-comment). Styles: questioning (probe), challenge (push on proof gaps), confirming (affirm lived insight — engage scores only unless force=true). Drafts are grounded in score, why[], and the original idea — never product names or CTAs, never auto-posted.

Is there a Chrome extension?

Yes — a minimal MV3 extension is available under extensions/antislop (load unpacked for dogfood). It scores selected text via the same evaluate API. You can also use the landing try-it widget, the dashboard AntiSlop tab, the embed script (/js/antislop-embed.js), or the open API directly.

Official package links: AntiSlop interactive demo

View AntiSlop with research tools