Neuroslop API

Connect detection, detailed analysis and humanization to your app or AI assistant through API and MCP.

Who it is for

Why this API

MCP server

Connect neuroslop to your AI agent via MCP — it checks text for slop, humanizes it and strips hidden characters on its own. Works in Claude Code, Cursor, Codex, OpenClaw.

How to connect

1. Free and keyless through the hosted HTTPS API; anti-abuse rate limits apply. Strips hidden characters: zero-width spaces, homoglyphs. Local, no key needed. Only neuroslop_humanize requires an API key.
2. Download neuroslop-mcp.mjs and add the config:

{
  "mcpServers": {
    "neuroslop": {
      "command": "node",
      "args": ["neuroslop-mcp.mjs"]
    }
  }
}

3. Replace the absolute path. NEUROSLOP_API_KEY is optional and is used only by neuroslop_humanize; neuroslop_check and neuroslop_clean work without a key.

Connect to your agent (MCP) Get a key in your account

Authentication

MCP checks use the free, rate-limited POST /api/check route. POST /v1/check requires a key and uses its units. GET /v1/health needs no key.

For /v1/check, /v1/rewrite and /v1/balance, send your key in a header: Authorization: Bearer <key> / X-API-Key: <key>.

Get a key two ways, no manual approval:

• On the site — your account → API: pick a pack, pay by card, SBP or USDT — the key is issued instantly.
• In the Telegram bot — @AITextChecker_bot → «🔌 API»: pay with Telegram Stars, the key arrives in the chat.

Pricing: units

POST/v1/check

Body: { "text": "...", "lang": "en" } (lang optional). Up to 20,000 characters.

Default limit for both endpoints: 20,000 UTF-16 code units, including spaces. A deployment may set a lower cap; GET /v1/health reports limits.maxInputChars, and too_long errors report max.

Examples below show selected response fields, not measured customer results. Scores and balances will differ.

curl -X POST https://neuroslop.net/v1/check \
  -H "Authorization: Bearer ns_YOURKEY" \
  -H "Content-Type: application/json" \
  -d '{"text":"The team utilizes this checklist to review each draft before publication.","lang":"en"}'

Response:

{
  "score": 100,
  "lang": "en",
  "ruleCoverage": "full",
  "scoreInterpretation": "style_rules_not_authorship",
  "applicability": { "status": "applicable", "eligible": true },
  "decision": {
    "outcome": "human_like",
    "action": "show_style_signal",
    "abstained": false,
    "confidence": null,
    "confidenceKind": "not_calibrated"
  },
  "assessment": {
    "outcome": "human_like",
    "authorshipProbability": null,
    "automaticRewrite": false
  },
  "remaining": 2999
}

Read applicability, decision and assessment before using a score. A 0–100 score is not an authorship probability; null means unavailable, not zero. lang describes the text, not the interface.

POST/v1/rewrite

Body: { "text": "...", "tier": "flash", "lang": "en" }. tier is flash (Standard, default) or pro (Premium). The input limit is reported by /v1/health.

curl -X POST https://neuroslop.net/v1/rewrite \
  -H "Authorization: Bearer ns_YOURKEY" \
  -H "Idempotency-Key: YOUR_UNIQUE_REQUEST_ID" \
  -H "Content-Type: application/json" \
  -d '{"text":"The team utilizes this checklist to review each draft before publication.","tier":"pro","lang":"en"}'

Response:

{
  "text": "Your revised paragraph appears here.",
  "tier": "pro",
  "scoreBefore": 100,
  "scoreAfter": 100,
  "qualityBasis": "editorial_review",
  "spent": 100,
  "remaining": 2900
}

qualityBasis describes the check, not the charge. Values include editorial_review, deterministic, bounded-source-witness, unchanged and unverified; do not treat this as a closed list. A higher detector score is not required for an approved rewrite.

{
  "text": "The team uses this checklist to review each draft before publication.",
  "tier": "pro",
  "scoreBefore": 100,
  "scoreAfter": 100,
  "unchanged": false,
  "reason": "local_edited",
  "qualityBasis": "deterministic",
  "engine": "local-rules",
  "editorialApproved": false,
  "billable": false,
  "spent": 0,
  "remaining": 3000
}

{
  "text": "Original text is returned here.",
  "tier": "pro",
  "scoreBefore": null,
  "scoreAfter": null,
  "unchanged": true,
  "reason": "local_unchanged",
  "qualityBasis": "unchanged",
  "spent": 0,
  "remaining": 3000
}

replayOnly: true

To retrieve a saved result without starting work, reuse the original Idempotency-Key, text, tier, lang and mode, adding replayOnly:true. A missing or expired receipt returns 409 idempotency_replay_unavailable, not a new rewrite. This does not prove whether the original request was charged. Do not automatically retry with a new key.

curl -X POST https://neuroslop.net/v1/rewrite \
  -H "Authorization: Bearer ns_YOURKEY" \
  -H "Idempotency-Key: YOUR_UNIQUE_REQUEST_ID" \
  -H "Content-Type: application/json" \
  -d '{"text":"The team utilizes this checklist to review each draft before publication.","tier":"pro","lang":"en","replayOnly":true}'

GET/v1/balance

Key quota left: { plan, quota, used, remaining, resetMonth, perMinute }.

quotaMode: "one_time", resetMonth: null

One-time unit pack. No monthly refill.

quotaMode: "monthly"

Monthly quota. Refills each calendar month.

GET/v1/health

Service liveness, no key needed: { "ok": true, "llm": true, "limits": { "maxInputChars": 20000 } }.

Errors

Handle error responses before using text. A timeout or 502 is not proof of the charge status: try saved-result recovery. Do not blindly retry authentication, quota or conflicting-key errors.