GPTZero
GPTZero · Tools & Agents
POST /v1/detectDeep-learning detector that flags portions of text likely generated by AI versus human, classifying content as entirely human, AI, or mixed.
At a glance
| Field | Value |
|---|---|
| Model id | gptzero |
| Model release date | - |
| Input modalities | Text, Document |
| Output modalities | Text |
| Context window | - |
| Weight precision | - |
| Features | ai_detection, classification |
| Native inference | No |
| New | No |
| Supported endpoints | POST /v1/detect, POST /v1/chat/completions, POST /v1/responses, POST /v1/messages, POST /v1beta/models/gptzero:generateContent |
| Alternate model ids | gptzero/detect |
Pricing
| Charge | Spec | Rate |
|---|---|---|
| Text Scan | per 1,000 words | $0.39 |
Example request
$ curl https://api.empiriolabs.ai/v1/detect \ > -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \ > -H 'Content-Type: application/json' \ > -d '{"model": "gptzero", "input": "The quick brown fox jumps over the lazy dog.", "scan_type": "ai_detection"}'
Parameters
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
input | string | no | - | Text to analyze |
scan_type | enum | no | "ai_detection" | AI Detection: detect AI-generated text. Bibliography: analyze citations and references. Sources / Fact Check: find supporting sources for claims in your text. · Allowed: ai_detection, bibliography, sources, fact_check |
model_version | string | no | "__latest__" | Select a specific GPTZero model version, or leave as latest for the newest rolling release. Choosing an explicit multilingual release overrides the separate multilingual flag. |
file_url | string | no | - | URL of a document to scan instead of pasting text. Supported: PDF, DOC, DOCX, RTF, ODT, TXT, MD (max 25 MB). The worker downloads the file and forwards it to GPTZero’s multipart files endpoint. Provide either input (text) or file_url, not both. |
version | string | no | "__latest__" | Canonical name for model_version. Pin a specific upstream snapshot, or leave as latest for the default rolling alias. |
multilingual | boolean | no | false | Use GPTZero’s multilingual detection path. Applies to text scans when Model Version is latest; file scans use whichever model version is selected. |
ai_threshold | number | no | 0.7 | Probability cutoff for flagging sentences as AI-generated. Lower = more sensitive. Range 0.10 to 0.99. · Range: 0.1 – 0.99 |
max_sentences | number | no | 15 | Maximum number of sentences shown in the per-sentence breakdown section. · Range: 1 – 50 |
source_sentence | string | no | - | Optional. The exact sentence or claim you want to find supporting sources for. If left blank, the full input text is used as the claim. (sources / fact_check scan only.) |
max_sources | number | no | 3 | Maximum number of sources to retrieve for each identified claim. (sources / fact_check scan only.) · Range: 1 – 10 |
confidence_threshold | number | no | 0.5 | Minimum confidence required for a source-match to be returned. Range 0 to 1. (sources / fact_check scan only.) · Range: 0 – 1 |
claim_extractor | string | no | "claimbuster" | Method used to extract claims from your text. claimbuster is faster; llm_claim_extractor is more thorough. (sources / fact_check scan only.) |
disable_formatting | boolean | no | false | Skip the EmpirioLabs Markdown report (per-sentence highlights / scan summary / sources list). The raw upstream GPTZero JSON response is returned. |
Notes
File limits
- Max 50 files per request
- Combined size cap: 15 MB
- Each document truncated to 50,000 characters
Supported formats
- PDF, DOC, DOCX, TXT, ODT
Machine-readable schema: GET https://api.empiriolabs.ai/v1/models/gptzero.