

AI Detector API
Identify AI-generated text with sentence-level probability scoring and confidence insights.
One API to detect, humanize, and verify AI content, so every word can prove its authorship. The same engine 2.1M writers and 12,000+ educators trust, now callable over the wire.
curl https://api.walterwrites.ai/v1/detect
-H "Authorization: Bearer YOUR_API_KEY" /
-H "Content-Type: application/json" /
-d '{
"text": "The rapid advancement of AI has transformed how we create content.",
"language": "en""id": "det_4k2p9x1m3",
"status": "completed",
"model": "detector-v3",
"created_at": "2026-04-24T13:45:00Z",
"output": {
"ai_probability": "0.87"
"words_processed": 12
"usage": { "credits_used": 12 }curl https://api.walterwrites.ai/v1/humanize
-H "Authorization: Bearer YOUR_API_KEY" /
-H "Content-Type: application/json" /
-d '{
"text": "Artificial intelligence is revolutionizing the way businesses approach content creation.",
"language": "en""id": "hum_8f3k2p9x1",
"status": "completed",
"model": "humanizer-v2",
"created_at": "2026-04-24T13:45:00Z",
"output": {
"text": "AI is changing how businesses think about creating content.",
"words_processed": 12
"usage": { "credits_used": 12curl https://api.walterwrites.ai/v1/plagiarism
-H "Authorization: Bearer YOUR_API_KEY" /
-H "Content-Type: application/json" /
-d '{
"text": "Innovation has always been the cornerstone of human progress and discovery.",
"language": "en""id": "plg_9x1m3k7p2",
"status": "completed",
"model": "plagiarism-v1",
"created_at": "2026-04-24T13:45:00Z",
"output": {
"originality_score": "0.13"
"words_processed": 11
"usage": { "credits_used": 11 }curl https://api.walterwrites.ai/v1/detect-image
-H "Authorization: Bearer YOUR_API_KEY" /
-H "Content-Type: application/json" /
-d '{
"image_url": "example.com/photo.jpg",
"language": "en""id": "img_2p9x1m3k7",
"status": "completed",
"model": "image-detector-v2",
"created_at": "2026-04-24T13:45:00Z",
"output": {
"ai_probability": "0.94"
"words_processed": 1
"usage": { "credits_used": 24 }curl https://api.walterwrites.ai/v1/grammar
-H "Authorization: Bearer YOUR_API_KEY" /
-H "Content-Type: application/json" /
-d '{
"text": "She dont like coffee but he go to the cafe every morning.",
"language": "en""id": "grm_1m3k7p2x9",
"status": "completed",
"model": "grammar-v2",
"created_at": "2026-04-24T13:45:00Z",
"output": {
"corrections_found": "2’’
"words_processed": 13
"usage": { "credits_used": 13 }Every Walter API is available through the same authentication, billing, and response structure. Pick the ones you need, add more later.


Identify AI-generated text with sentence-level probability scoring and confidence insights.


Refine AI-generated text into natural, human-sounding output that reads authentically.


Detect AI-generated and cloned voice in audio with speaker-level confidence scoring.


Scan text against billions of web sources and academic databases for originality matches.


Catch grammar, spelling, and style issues with context-aware, sentence-level suggestions.


Detect AI-generated and manipulated images across major models with source attribution.
Every endpoint draws from a single monthly credit balance. Start on the free tier, then pick the plan that fits your volume.
Need enterprise pricing?
Dedicated capacity, region pinning, and Zero-Retention mode for high-volume teams.
Production teams need more than an endpoint. Walter is built with the privacy, reliability, and compliance practices enterprise buyers look for.
Built for high availability, with infrastructure across US, EU, and AP regions.
Learn more


We document how each endpoint works, so behavior stays clear and predictable.
Learn more


SOC 2 Type I, with privacy practices aligned to GDPR and CCPA. Audit logging and scoped API keys.
Learn more

Token-based authentication, with signed webhooks and idempotency keys on writes. Rate limits scale with your plan.
Learn moreFirst-party SDKs map the full OpenAPI spec, support streaming, and ship with retries built in.
$ pip install walter from walter import Client client = Client(api_key=“ww_…”) result = client.humanize(“Your text”)
$ npm install walter import { Client } from “walter”; const client = new Client( { apiKey: “ww_…” }); const result = await client.detect(“Your text”);
$ gem install walter require “walter” client = Walter::Client.new(api_key:“ww_…”) result = client.detect(“Your text”)
$ go get github.com/walter/walter-go import “github.com/walter/walter-go” client := walter.=NewClient(“ww_…”) result, _ := client.Detect(“Your text”)
$ composer require walter/walter use WalterClient; $client = new Client“ww_…”); $result = $client->detect(“Your text”)
// Maven dependency: ai.walter:walter-java import ai.walter.Client; Client client = new Client (“ww_…”) Result result = client.detect(“Your text”)
$ dotnet add package walter using Walter; var client = new Client=“ww_…”) var result = await client.DetectAsync(“Your text”)
Teams use Walter APIs to verify, refine, and protect content at every stage of the workflow.
Flag AI-generated and synthetic content before it reaches your users. Run posts, uploads, and submissions through Detector and Image Detector APIs to keep your platform authentic.
Help educators assess originality at scale. Combine Detection and Plagiarism Checker to surface AI-assisted work and potential source matches in a single response.
Speed up content review with automated humanization, grammar checking, and originality scoring. Ship cleaner drafts with every call tracked in one dashboard.
Detect AI-generated images before they spread. Flag manipulated profile photos, fake product listings, and altered uploads in real time.
Power writing tools with humanization and grammar built in. Give your users natural-sounding drafts and polished copy from a single integration.
Verify originality before publication. Scan manuscripts and articles against billions of sources to catch issues early in the editorial cycle.
Everything you need to know about the Walter API.
One key gives you the full Walter suite over HTTP: detect AI-generated text, images, and voice, humanize AI drafts into natural writing, check plagiarism, and run grammar checks. Every endpoint shares the same authentication, request, and response format, so you can start with one and add others with no new integration work.
Detection returns a calibrated probability, scored sentence by sentence rather than a single yes or no. Treat it as a signal to review, not a verdict, and set a threshold that fits your risk tolerance. We report results on public benchmarks and do not train on the test set, so the numbers reflect real generalization.
Yes, by anyone determined enough. The 2023 Sadasivan et al. paper “Can AI-Generated Text Be Reliably Detected?” showed every detector has theoretical evasion paths through paraphrasing. That is exactly why we ship a humanizer alongside the detector: they are two sides of the same problem. Use the score as one input in a broader review, not as proof.
No. We do not train our models on the text, images, or audio you send through the API. Your content is processed only to return your result, not to train or improve our models. See the docs for full data-handling details.
Usage is credit-based. Every API draws from one monthly credit balance, and text, images, and voice each consume credits at their own rate. You start with a free tier, then choose a plan that matches your expected volume. One balance covers every endpoint, so there are no separate budgets to manage per API.
Sign up, create a key in your dashboard, and pass it as a Bearer token in the Authorization header. Your first calls are free with no credit card required, and the same key works across every Walter endpoint. Full setup and code samples are in the docs.