How accurate is the AI detector?

The AI detector returns a headline 0–100 score plus sub-scores from multiple engines, which is more robust than any single checker. Accuracy improves with length — feed it at least ~200 words. No AI detector is perfect; treat the score as a strong signal for routing and review, not an absolute verdict.

How does AI detector pricing work?

Credit-based: 1 credit equals $0.001 USD. The AI detector bills per word at one-tenth the humanize rate, with a 50-word floor — short samples are charged as 50 words. See the current per-1,000-words rate in the pricing table on this page. Failed requests are refunded automatically.

What does the AI detector score mean?

A single number from 0 to 100: roughly, under 50 reads as human-written, 50–60 is possible AI, and over 60 is likely AI. The AI detector also returns each engine sub-score so you can see whether the engines agree or split on a given text.

How long can the input text be?

The AI detector accepts up to 30,000 words per request. For the most reliable score, give it at least ~200 words — very short snippets carry less signal, though the AI detector still returns a result.

How do I get the result back?

The AI detector is async. POST to /api/v1/detect and you get a task_id; poll /api/v1/tasks/:id until status is completed, then read the detection object from the output. Typical checks finish in a few seconds; polling adds no extra charge.

Where is the full AI detector API reference?

The complete parameter list, score interpretation, error codes, and runnable cURL / Python / Node examples for the AI detector live at /docs/ai-text-detector.

Is AI Text Detector API free or paid?

Paid. Completed generations are charged at the rates in this page's pricing table. Eligible new accounts may receive promotional test credits; availability and amount are shown at signup or in the dashboard. Technical or provider failures are refunded automatically, except where this page documents a charged output-moderation block.

Does AI Text Detector API require a monthly subscription?

No. Pricing is pay-as-you-go: $1 buys 1,000 credits and credits never expire. Cost scales with what you generate rather than with a recurring fee.

How do I get an AI Text Detector API key?

Sign up and create a key in the dashboard, no card required. Pass it as a Bearer token in the Authorization header.

AI Text Detector API — Check Text for AI With One Call

This AI detector scores any text 0–100 for AI authorship and returns a per-engine breakdown. Send text, poll the task, and the AI detector tells you how likely the writing is machine-generated — over the same async /api/v1 contract you already use.

Highlights: 8 detection engines · one aggregate score
Best for: Screening submissions, QA on drafts
Input: Text to analyze
Output: AI-likelihood score + per-engine breakdown

Input

Text*

Under 30,000 words · at least ~200 words gives the most accurate score

Result

Try one of these prompts

Pricing

Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.

First test cost

$0.08989 credits

One sample at the cheapest tier (1,000 words).

Testing budget guide

Add credits

$10

≈ 112 tests

$50

≈ 561 tests

$100

≈ 1123 tests

Any text: $0.089; Price · 1,000 words

What is the AI Text Detector API?

AI Text Detector scores any passage from 0 to 100 for how likely it is to be AI-generated, aggregating several third-party detection engines into one headline result plus the per-engine sub-scores behind it. One async call handles texts of under 30,000 words — around 200 words or more gives the strongest signal — and every request is billed per word of the text you submit, with a 50-word floor.

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

AI detector scoring a batch of user submissions

Screen submissions with the AI detector

Platforms that take user writing — essays, applications, reviews, forum posts — run each submission through the AI detector to flag machine-generated content before it lands. It returns a single 0–100 score plus a per-engine breakdown, so a moderation queue can sort by confidence and route only the borderline cases to a human. The async task contract means a pipeline calls the AI detector, gets a task_id, and moves on.

Score some text

AI detector verifying a freelance writing deliverable

Verify freelance and agency deliverables

When you pay for human writing, the AI detector confirms you got it. Paste the delivered copy and it reports how AI-like it reads, with sub-scores from multiple third-party engines so the verdict is not a single black box. Per-word billing keeps spot-checks cheap: the AI detector costs a fraction of the humanize rate, with a small 50-word floor for short samples.

AI detector re-checking humanized text in a QA loop

QA your own humanized output

Pair the AI detector with the humanizer to close the loop: rewrite a draft, then run the AI detector to confirm the score dropped before you publish. Because it aggregates several engines into one result, you get a more robust read than any single checker. Submit, poll, read the score — same task lifecycle as every reApi model.

Why reAPI for AI Text Detector

A multi-engine AI detector in one call

It aggregates several third-party detection engines into a single 0–100 score and exposes each engine sub-score alongside it. One request to /api/v1/detect, one task_id, one result that is harder to game than any single checker — no need to fan out to multiple vendors and reconcile their answers yourself.

Per-word pricing, a tenth of humanize

1 credit equals $0.001 USD. The AI detector bills per word at one-tenth the humanize rate with a 50-word floor, so screening costs scale with the text you check — no seat fees, no monthly minimum. Failed requests are refunded automatically; you only pay for scores that come back.

Async-first for production stacks

Submit text, get a task_id, poll until done — usually a few seconds. The AI detector fits the same task plumbing as every other reApi model, so a moderation or QA service drops it in without new infrastructure. One JSON shape, one Bearer key, one credit ledger.

AI detector vs a single-engine checker

Why call this AI detector instead of wiring up one detection vendor directly? It comes down to coverage, billing, and how much integration you own.

Capability

AI detector on reApi

Single-engine checker

Engine coverage

Aggregates several engines into one score plus sub-scores — harder to game.

One model; a text tuned to beat that single engine slips through.

Result shape

Headline 0–100 plus per-engine breakdown in one response.

A single number with no second opinion.

Billing

Per word, 50-word floor, refunded on failure — costs scale with text checked.

Often per-document or monthly seats regardless of volume.

Integration

One /api/v1/detect task endpoint, same auth and credit ledger as your other models.

A separate vendor account, SDK, and billing relationship to maintain.

Input length

Up to 30,000 words per call.

Caps and per-call limits vary by vendor.

Pairs with humanize

Same platform as the humanizer for a detect → rewrite → detect QA loop.

Detection only; you source humanizing separately.

Comparison reflects publicly documented behavior at the time of writing. Detection accuracy varies by text and engine and changes over time.

Integrate the AI detector in three steps

step 01
Create an API key
Sign up at reApi and grab a key. Free signup credits cover plenty of words through the AI detector — enough to validate the scores against text you already know the origin of.
Open
step 02
Submit text to check
POST to /api/v1/detect with model: ai-text-detector and your text. The AI detector returns a task_id immediately so your handler is free to move on.
Open
step 03
Poll for the score
GET /api/v1/tasks/:id until status is completed. The response carries detection — the 0–100 AI-likelihood plus the per-engine sub-scores.
Open

docs/api/ai-text-detector

API reference

Drop-in code and the full parameter table.

View full reference

curl -X POST https://reapi.ai/api/v1/detect \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ai-text-detector",
    "text": "Paste the text you want to check for AI authorship here. At least 200 words gives the most accurate score."
  }'

Parameter validation failed

Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.

Authentication or balance issue

401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.

Content or material rejected

Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.

Task failed or timed out

A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.

Frequently asked questions

Common questions about this model.

The AI detector scores any text 0–100 for how likely it is AI-generated, aggregating several third-party detection engines into one result. On reApi the AI detector ships as model id ai-text-detector on the /api/v1/detect endpoint, returns the verdict as a detection object over an async task, and is billed per word. See the current rate in the pricing table on this page.