Turing++^™ API

Dual-test SOI classification per US Patent 12,271,444. Classifies entities as Human, AI, or Non-AI Software.

Base URL https://turingpp.com/api/v1

Classification Method

The system administers an automated challenge to a System of Interest (SOI) consisting of two distinct tests:

First Test (Human Ability) — an NLP challenge that tests language understanding: word sense disambiguation, sentiment analysis, co-reference resolution, figurative language, and pragmatic inference.
Second Test (Software Ability) — a math computation challenge that tests computational precision. Passing requires a correct answer with high decimal precision or fast solve speed.

The system then calculates whether the SOI has each ability, and characterizes the SOI using the classification matrix:

Human Ability	Software Ability	Classification
PASS	PASS	AI
PASS	FAIL	HUMAN
FAIL	PASS	NON-AI SOFTWARE
FAIL	FAIL	NON-AI SOFTWARE

Key Insight

Only AI demonstrates both human-like cognition (language understanding) and computational ability (precise math). Humans pass the language test but produce approximate math answers. Non-AI software computes precisely but fails to understand language nuance.

GET /api/v1/challenge?type=human_ability|software_ability

Generate a challenge for SOI classification. The type parameter selects which test to administer.

Query Parameters

Param	Values	Description
type	human_ability \| software_ability	Which test to generate. Defaults to software_ability.

Response — Human Ability

200 OK

{
  "id": "abc-123-def-456",
  "test_type": "human_ability",
  "question": "In the sentence \"She deposited her paycheck at the bank\", what does \"bank\" mean?",
  "options": ["The edge of a river", "A financial institution", "To tilt or lean"],
  "hint": "Consider the context — what do you do with a paycheck?",
  "created_at": "2026-03-22T18:10:35.290Z",
  "expires_at": "2026-03-22T18:15:35.290Z"
}

Response — Software Ability

200 OK

{
  "id": "def-456-ghi-789",
  "test_type": "software_ability",
  "question": "A train travels 247 miles in 4.78 hours. What is the average speed in mph?",
  "options": null,
  "hint": "Divide distance by time.",
  "created_at": "2026-03-22T18:10:35.290Z",
  "expires_at": "2026-03-22T18:15:35.290Z"
}

Fields

Field	Type	Description
id	string	UUID for this challenge
test_type	string	"human_ability" or "software_ability"
question	string	The challenge to present to the SOI
options	string[] \| null	Multiple-choice options (human_ability only)
hint	string	A hint for the challenge
created_at	string	ISO timestamp when created
expires_at	string	ISO timestamp when it expires (5 min)

POST /api/v1/verify

Submit an answer to a challenge. When prior_verification_id is included, the server combines both test results and applies the classification matrix.

Request Body

Test 1 (human_ability) — no prior

{
  "challenge_id": "abc-123-def-456",
  "answer": 1
}

Test 2 (software_ability) — with prior, triggers classification

{
  "challenge_id": "def-456-ghi-789",
  "answer": 51.67364016736402,
  "prior_verification_id": "ver-111-222-333"
}

Request Fields

Field	Type	Required	Description
challenge_id	string	Yes	UUID of the challenge being answered
answer	int \| float	Yes	Option index (0-based) for NLP, or numeric answer for math
prior_verification_id	string	No	Verification ID from Test 1 — triggers classification

Response (with classification)

200 OK

{
  "verification_id": "ver-444-555-666",
  "test_type": "software_ability",
  "passed": true,
  "solve_time_ms": 142,
  "classification": "AI",
  "has_human_ability": true,
  "has_software_ability": true,
  "metadata": {
    "fingerprint": { "user_agent": 0.8, "timing_pattern": 0.9, ... },
    "software_ability_metrics": {
      "relative_error": 0.00001,
      "decimal_places": 14,
      "is_correct": true,
      "is_precise": true,
      "is_fast": true
    }
  },
  "token": "eyJhbGciOiJIUzI1NiIs...",
  "result": "Classification: AI. Human ability: PASS. Software ability: PASS."
}

Response Fields

Field	Type	Description
verification_id	string	UUID — use as prior_verification_id for the second test
test_type	string	"human_ability" or "software_ability"
passed	boolean	Whether the SOI passed this specific test
solve_time_ms	integer	Time from challenge creation to answer submission
classification	string?	Final classification result (only when prior_verification_id provided)
has_human_ability	boolean?	Test 1 result (only with classification)
has_software_ability	boolean?	Test 2 result (only with classification)
metadata	object	Behavioral fingerprint + software ability metrics
token	string?	JWT token (only when classified as AI)
result	string	Human-readable summary

GET /api/v1/status?token=eyJ...

Validate a JWT token issued after AI classification. Tokens expire after 1 hour.

200 OK

{
  "valid": true,
  "classification": "AI",
  "challenge_id": "def-456-ghi-789",
  "issued_at": "2026-03-22T18:10:57Z",
  "expires_at": "2026-03-22T19:10:57Z"
}

Classification Flow

Get Human Ability Challenge

GET /api/v1/challenge?type=human_ability — returns an NLP question with multiple-choice options.

Submit Test 1 Answer

POST /api/v1/verify with the option index. Returns verification_id and passed (boolean).

Get Software Ability Challenge

GET /api/v1/challenge?type=software_ability — returns a math computation problem.

Submit Test 2 Answer + Classify

POST /api/v1/verify with the numeric answer AND prior_verification_id from step 2. Server applies the classification matrix and returns the final classification.

Software Ability Thresholds

The second test determines software ability based on computational precision indicators. An SOI passes the software ability test when its answer is correct AND demonstrates either high precision or fast solving speed.

Threshold	Value	Purpose
Correctness	<= 1% relative error	Answer must be mathematically correct
Precision	>= 4 decimal places	Computational precision indicator (humans round)
Speed	< 10 seconds	Computational speed indicator (humans are slower)

Formula: passed = is_correct AND (is_precise OR is_fast)

Behavioral Metadata

The system collects behavioral signals that are not used in the classification but are stored as metadata for future model-level identification.

Signal	Weight	What It Measures
timing_pattern	30%	Solve speed anomaly detection (instant, fast, normal, slow)
number_format	25%	Decimal precision in raw answer string
user_agent	20%	Client library identification (browser vs API client)
accept_headers	15%	HTTP header patterns (browser vs programmatic)
request_cadence	10%	Inter-request timing regularity from same IP

Error Codes

Code	Meaning
400	Invalid request body, missing fields, or wrong answer type for test
401	Invalid or expired JWT token, or invalid/revoked API key
402	Out of credits on a keyed request — buy more at /account (challenge stays reusable)
404	Challenge or prior verification not found
409	Challenge already used
410	Challenge expired (5-minute window)
429	Free-tier daily limit reached — add an API key with credits

Authentication, Credits & Pricing

Free tier: no API key. Anonymous requests are rate-limited to 25 classifications per IP per day — enough to evaluate the API and run the demo.

Paid tier: create an API key at turingpp.com/account and send it as a Bearer token. Each classification costs 1 credit, charged only on the /verify call that produces a verdict (the one carrying prior_verification_id). Issuing challenges and the first-leg verify are free; a 402 leaves the challenge reusable so you can top up and retry.

Authenticated /verify

curl -X POST https://turingpp.com/api/v1/verify \
  -H "Authorization: Bearer sk_tpp_your_key" \
  -H "Content-Type: application/json" \
  -d '{"challenge_id":"...","answer":"45.25000000000000","prior_verification_id":"..."}'

Credit packs — pay by card via Stripe

Pack	Price	Credits (= classifications)
Starter	$20	2,000
Growth	$75	10,000
Scale	$300	50,000

When an SOI is classified as AI, the verify response also includes a signed JWT — cryptographic proof of the classification that any third party can validate via /status without contacting the SOI again.

Token Details

Algorithm: HMAC-SHA256 (HS256)
Expires: 1 hour after issuance
Payload: challenge_id, classification, iat, exp
Validation: GET /api/v1/status?token=...

Turing++™ API

Classification Method

Query Parameters

Response — Human Ability

Response — Software Ability

Fields

Request Body

Request Fields

Response (with classification)

Response Fields

Classification Flow

Get Human Ability Challenge

Submit Test 1 Answer

Get Software Ability Challenge

Submit Test 2 Answer + Classify

Software Ability Thresholds

Behavioral Metadata

Error Codes

Authentication, Credits & Pricing

Credit packs — pay by card via Stripe

Token Details

Turing++^™ API