Membrace will be joining forces with Sellesta later this year! Stay tuned for updates.

Content
Moderation API

Checks video, text, image, audio for any unwanted content

Our solutions can flexibly adapt
to your business’s values

Images

Check the quality and appropriateness of images to make your service even more attractive to users.
Show more

Text

Detect unpleasant or rude communication or mentions of prohibited products and services.
Show more

Video

Check videos for prohibited scenes and content that might interfere with a pleasant viewing.
Show more

Audio

Identify unwanted content in audio tracks, check the sound quality and signal.
Show more

Solution can detect the following classes

Spam

Content that isn't useful and aims to lure the reader to some internet resource.

Clickbait

A message that excites and arouses curiosity but doesn't give complete information.

Obscene

Rude expressions, including those censored with asterisks or dots.

Profanity

Insults addressed at any people or material objects.

Offensive content

Insults addressed at any people or material objects.

Erotica

Descriptions of sex or someone's appearance that are informational
in nature and don't mention gross physiological details.

Sex toys

Any mention of sex toys and the experience of using them, and reviews without physiological details.

Violence

Shocking or disgusting text. Calls for violence and cruelty. Encouraging hatred on grounds of gender, race, nationality, or association with any social group

Drugs

The mention of drugs in any context, including their promotion, descriptions of manufacture methods, and the use and sale of prohibited narcotic substances.

Quickstart: Use the Content
Moderator client API

Request
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
import requests
 
token = 'abacaba'  # set real token here
api_url = 'https://api.membrace.ai/v1/item/process'
headers = {'Authorization': token}
 
# Text request
text = 'Hello world'
resp = requests.post(
    api_url,
    headers=headers,
    json={
        'request_id': '1',
        'project_id': 'text',
        'project_version': '1',
        'content': {
            'text': text
        }
    }
)
resp.raise_for_status()
print(resp.json())
Response
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
{
  "request_id": "1",
  "item_id": "safdhgjkl",
  "project_id": <project_id>,
  "project_version": "1",
  "status": "SUCCESS",
  "result": {
      "alcohol": false,
      "clean_text": true,
      "clickbait": false,
      "drugs": false,
      "erotica": false,
      "explicit": false,
      "insult": false,
      "insult_light": false,
      "insult_strong": false,
      "obscene_light": false,
      "obscene_strong": false,
      "personal_data": false,
      "swearing": false
  }
}
Request
1
2
3
4
5
6
7
8
9
10
11
12
curl -X POST \
     -H 'Authorization: <token>' \
     -H 'Content-Type: application/JSON' \
     -d '{
      "request_id": "1",
      "project_id": <project_id>,
      "project_version": "1",
      "content": {
        "text": "Example message"
      }
    }' \
https://api.membrace.ai/v1/item/process
Response
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
{
  "request_id": "1",
  "item_id": "safdhgjkl",
  "project_id": <project_id>,
  "project_version": "1",
  "status": "SUCCESS",
  "result": {
      "alcohol": false,
      "clean_text": true,
      "clickbait": false,
      "drugs": false,
      "erotica": false,
      "explicit": false,
      "insult": false,
      "insult_light": false,
      "insult_strong": false,
      "obscene_light": false,
      "obscene_strong": false,
      "personal_data": false,
      "swearing": false
  }
}
Have questions? Talk to an expert

Why working with Membrace?

Large and responsive crowd

We have over 2M workers distributed over the globe, so manual checks are available 24/7

Scalability

We can process any volume of data, including spikes caused by spam attacks, breaking news, flash mobs, or organic growth

Flexibility

Customizable moderation classes for brand-appropriate responses

Continuous fine-tuning

Our crowd force can handle any granularity of detail, and our models continue to learn from incoming user data

High accuracy
in unique cases

We apply a multitude of QA methods to both our human-in-the-loop moderators and our ML models

Real-time moderation

For rapidly growing, large-scale, and fast-paced businesses

Ready to get started?

Сookie settings
By clicking «Accept All», you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Try it on Your
Dataset for free
Try it on your dataset for free