Classify and filter

Hush is a text classification API that detects harmful and toxic content in any text. One API call is all it takes — integrate in minutes and moderate at any scale.

See it in action Read the Docs
50
ms latency
99%
accuracy
1M+
calls/day
How it works

Built for production, from day one

Three properties that make Hush different from basic keyword filters.

Low Latency

Classifications are returned in milliseconds, making it suitable for real-time content pipelines at any scale.

Context Aware

Goes beyond keyword matching to understand the intent and tone behind any piece of text.

Easy to Integrate

A single REST API call is all it takes. Works with any language or framework that can send an HTTP request.

Live Playground

Give it a try!

Type any text and see Hush classify it in real time. Paste a comment, message, or any string to see the API at work.

Real-time classification via the live API
Integrate in minutes with a single API call
Transparent, usage-based pricing

Plans built for every scale

Pick the plan that fits your traffic today and scale seamlessly as you grow. Same API, consistent latency, better support tiers.

Starter

Free

Forever

$0

Kick the tyres with generous limits for prototypes and QA.

  • 100 free tokens each month
  • Community support
  • Live playground access
Activate Free

Growth

Basic

Popular

$9.99

Great for pilots and early production workloads.

  • 10K tokens monthly
  • Email support
  • Latest model release
Choose Basic

Scale

Pro

Full access

$19.99

Unlimited moderation plus priority response times.

  • Unlimited tokens
  • Priority support
  • Latest model release
  • Model selection
Upgrade to Pro

API Documentation

Everything you need to integrate Hush into your platform. Our REST API is simple, predictable, and incredibly fast.

Getting Started

  • Base URL
  • Error Codes

Endpoints

  • POST /v1/predict
  • GET /health

Base URL

https://api.openproject.co.zw
POST /models/hush-preview:03-2026/v1/predict

Analyze a text message for toxicity. Returns a boolean classification indicating whether the message is toxic or safe.

Request Headers
"Content-Type": "application/json"
Request Body
{
  "text": "You are doing a great job!"
}
Response 200 OK
{
  "is_toxic": false,
  "prediction": "non-toxic"
}
Error Responses
400 — Missing or empty "text" field
{ "error": "No text provided" }
401 — Missing or invalid API key
{ "message": "API key not found" }
402 — Insufficient tokens
{ "error": "Insufficient tokens", "tokens_remaining": 4 }
429 — Rate limit exceeded (5 req/min)
{ "error": "Rate limit exceeded" }
GET /health

Check whether the Hush API is online and running.

Response 200 OK
{
  "status": "healthy",
  "service": "hush"
}

Integration Examples

python 
import requests

url     = "https://api.openproject.co.zw/models/hush-preview:03-2026/v1/predict"
headers = { "Authorization": "Bearer YOUR_API_KEY" }
data    = { "text": "This community is amazing!" }

response = requests.post(url, headers=headers, json=data)
result   = response.json()

if result['is_toxic']:
    print("Message flagged.")
else:
    print("Message is safe to send!")
javascript 
const response = await fetch(
    "https://api.openproject.co.zw/models/hush-preview:03-2026/v1/predict",
    {
        method:  "POST",
        headers: {
            "Content-Type":  "application/json",
            "Authorization": "Bearer YOUR_API_KEY"
        },
        body:    JSON.stringify({ text: "This community is amazing!" })
    }
);

const result = await response.json();

if (result.is_toxic) {
    console.log("Message flagged.");
} else {
    console.log("Message is safe to send!");
}
curl 
curl -X POST https://api.openproject.co.zw/models/hush-preview:03-2026/v1/predict \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"text": "This community is amazing!"}'