Jev Bulk Classifier — Text & Image Labels, Ratings, Yes/No

Turn a list of texts or images into structured columns. Give it your own labels — support ticket queues, lead grades, sentiment, spam, intent — and every text or image comes back with one typed answer per question, plus a confidence score you can filter on.

No prompt engineering, no JSON parsing, no free-form text to clean up. You define the options; the answer is always one of them.

What you get

For every text, image or record, one dataset row with:

• Choice questions — the selected label (department: "billing"), a confidence score, and the probability of every option
• Score questions — a numeric rating against your own ordered levels (frustration: 1.68), the nearest level's label, confidence, and per-level probabilities
• Yes/no questions — a plain true/false plus the underlying probability, so you can set your own threshold
• The original text and image URLs (and the full source record when you read from a dataset), so results stay joinable
• Export to JSON, CSV, or Google Sheets directly from the Apify console

Ask up to 10 questions at once — they are answered in a single pass per text, so adding questions costs you nothing extra in runtime.

Use cases

• Support ticket triage — route each message to the right team and flag the urgent ones
• Review and comment analysis — sentiment, topic, and "is this a complaint?" across thousands of reviews at once
• Lead qualification — score scraped profiles or form submissions against your own ICP criteria
• Content moderation and spam filtering — flag off-topic, promotional, or abusive text and images with a tunable confidence threshold
• Image tagging — sort product photos, listing images or social posts into your own categories, or flag the ones that break your rules
• Cleaning up scraper output — chain it onto any other actor's dataset to label or filter the rows before they reach your CRM or warehouse

How to use

1. Paste your texts into Texts and/or image links into Image URLs, one per line — or put a Dataset ID from a previous run to label an existing dataset in place
2. When reading from a dataset, list the Fields to evaluate (e.g. title, body) so only the relevant text is sent, and the Image fields (e.g. displayUrl, images) to evaluate a record's images together with its text
3. Define your Questions. Each one needs:
   • key — the column name in the output
   • type — choice (pick one option), score (rate against ordered levels), or noul (yes/no)
   • instructions — what you want decided
   • criteria — the options for choice, the ordered levels for score; optional for noul
4. Set Max Items (default 100; set 0 for no limit) and Concurrency
5. Run the actor — results appear in the Dataset tab

Question types

[
  {
    "key": "department",
    "type": "choice",
    "instructions": "Which team should handle this message?",
    "criteria": {
      "billing": "Payment or subscription issues",
      "technical": "Bugs or integration problems",
      "sales": "Pricing or account questions"
    }
  },
  {
    "key": "frustration",
    "type": "score",
    "instructions": "How frustrated the customer appears",
    "criteria": [
      "Calm, just stating facts",
      "Frustrated but civil",
      "Very angry, strong language"
    ]
  },
  {
    "key": "is_urgent",
    "type": "noul",
    "instructions": "The message conveys urgency or time-sensitivity"
  }
]

A choice question also accepts a plain list of labels (["positive", "neutral", "negative"]) when the labels speak for themselves.

Output format

Each dataset record:

{
  "itemIndex": 0,
  "text": "Hi, I've been trying to connect my Stripe account for 3 days and the integration keeps failing. I'm losing sales. Please help ASAP.",
  "department": "technical",
  "department_confidence": 0.8,
  "department_probabilities": { "technical": 0.87, "billing": 0.13, "sales": 0 },
  "frustration": 1,
  "frustration_label": "Frustrated but civil",
  "frustration_confidence": 1,
  "frustration_probabilities": { "0": 0, "1": 1, "2": 0 },
  "is_urgent": true,
  "is_urgent_probability": 1
}

If a text cannot be evaluated, the row is still written with an error field so nothing silently disappears — the rest of the run continues.

Notes

• Each text is truncated to 8,000 characters before it is evaluated
• Images can be JPEG, PNG, WebP or GIF (up to 20 MB). They are resized automatically, and up to 4 images per dataset record are evaluated together
• If an image cannot be downloaded, the row gets an imageErrors field; the item is still evaluated on whatever text and images remain
• Confidence is separate from probability: use <key>_confidence to decide whether to act automatically or send a row to a human
• Texts longer than a few paragraphs work best when you narrow them down with Fields to evaluate