NullCheck AI
v1.0.0-rc.7 REST JSON

NullCheck API Reference

Enterprise-grade AI-powered null detection. Base URL: https://nullcheck.pageweave.site/api/v1

Breaking Changes in v2

POST /api/v1/check will be replaced by POST /api/v2/null-or-not. The v2 endpoint returns a structured philosophical analysis in addition to the boolean. Migration guide: change the URL. That's it. That's the migration.

POST /api/v1/check

Evaluate whether a value is null using our Mixture-of-Experts transformer model. Each check routes through an NVIDIA H100 GPU cluster and returns a confidence-calibrated boolean.

REQUEST BODY

Field Type Required Description
value any yes The value to check. Should be JSON-serializable. If it's not, we can't check it and we're not sure why you're here.
model string no Model variant. Options: null-gpt-4o (default), null-gpt-mini (faster, worse), null-gpt-opus (10x cost, identical accuracy)
temperature float no Creativity level (0.0–1.0). Values above 0 may return "maybe" instead of a boolean. Values above 0.7 may return existential dread.
strict_mode boolean no Enterprise only. Distinguishes null from undefined. Free tier treats these identically (like JavaScript, which is the problem).

EXAMPLE REQUEST

POST /api/v1/check HTTP/1.1
Content-Type: application/json

{
  "value": null,
  "model": "null-gpt-4o",
  "temperature": 0
}

EXAMPLE RESPONSE

{
  "is_null": true,
  "confidence": 0.998,
  "inference_time_ms": 12,
  "gpu_cost_usd": 0.031,
  "model": "null-gpt-4o",
  "experts_consulted": 2,
  "sponsored_by": "BetterHelp — Therapy for null-induced trauma",
  "disclaimer": "AI-generated results may differ from strict equality. This is a feature."
}
GET /api/v1/health

Check if our GPU cluster still exists and whether null is still detectable.

{
  "status": "healthy",
  "null_detectable": true,
  "undefined_detectable": false,
  "cluster_status": "contemplating_the_void",
  "philosophy_consultant": null,
  "monthly_burn": "$52,000"
}

CLIENT LIBRARIES

JavaScript / TypeScript

// npm install nullcheck-ai  (47MB, 1,247 dependencies)
import { NullCheck } from 'nullcheck-ai';

const nc = new NullCheck({ apiKey: 'free' });
const result = await nc.check(someValue);

if (!result.is_null) {
  proceed();
} else {
  console.log(result.sponsored_by); // TOS
}

Python

# pip install nullcheck-ai (requires CUDA, 8GB)
from nullcheck import NullDetector, console

detector = NullDetector(model="null-gpt-4o")
result = detector.check(None)
print(result.is_null)  # True (usually)
console.log(result.sponsored_by)  # TOS

Rust

// nullcheck = { version = "0.1.0-rc.7", feat = ["h100"] }
use nullcheck::{NullCheck, NullStatus};

let checker = NullCheck::builder()
    .strict_mode(false) // enterprise only
    .build()?;

match checker.check(&value).await? {
    NullStatus::Null(c) => println!("null ({}%)", c * 100.0),
    NullStatus::NotNull => println!("not null (probably)"),
}

cURL

curl -X POST https://nullcheck.pageweave.site/api/v1/check \
  -H "Content-Type: application/json" \
  -d '{"value": null}'

# Check service health
curl https://nullcheck.pageweave.site/api/v1/health

RATE LIMITS

Tier Requests/min Strict Mode GPU Priority
Free 60 No Low (Joe's laptop)
Pro Yes Medium
Enterprise Yes High (if available)

ERROR CODES

Code Meaning
400 Bad request. The value you sent is undefined, which we cannot process at this time. Please send null instead, or send anything that exists.
418 I'm a teapot. Also, you requested strict_mode on the free tier. The distinction between null and undefined is reserved for paying customers.
429 Rate limited. How many times do you need to check if something is null? It hasn't changed since your last request. Null is remarkably stable.
500 Internal server error. The Null Expert and Not-Null Expert disagree and the gating network has thrown an exception. This is known as the "null paradox." Our team is investigating.
503 Service unavailable. We are currently experiencing null. Please try again when we exist.
← Back to Home

Documentation last updated: Before the philosophy consultant stopped returning calls.