Classify Any URL
Any Taxonomy
One API Call
URL-first, taxonomy-agnostic classification. We scrape, extract, and classify in one hop — IAB 2.2 or your own. Raw text works too.
Built for classification at scale
Six primitives that make up every request. Understand them once, predict every response.
URL in, labels out
Send a URL, we scrape, extract the body, and classify. One hop, SSRF-safe, 5MB cap. Raw text is accepted too — skip the network when you already have the content.
Taxonomy-agnostic
IAB 2.2, sentiment, and toxicity ship today. One key, one response shape across every taxonomy. More models on the way.
Better economics
At 150k classifications a month we cost under a third of Klazify. At 500k we're a fifth. Rate cards stay on-page — no per-token math, no sales call to see a number.
Built for three jobs
Contextual ad targeting with IAB 2.2. Brand safety and moderation with toxicity. Content routing and CRM enrichment with sentiment. One API, the patterns your peers already expect.
Fine-tuned, not zero-shot
The iab-2.2 model is a fine-tuned DeBERTa-v3-large classifier, trained on web content we collected and labeled ourselves. Not a prompt wrapped around an LLM. Deterministic outputs, cheaper per request, and the taxonomy ships with the model.
Structured output
Ranked categories, confidence floats, taxonomy paths. Every response is the same shape across every model. No prompt engineering, no regex on prose.
One request from text to taxonomy
Language SDK or raw cURL — the contract is the same. Send text or a URL, receive ranked categories with confidence scores. No prompt engineering, no schemas to maintain.
# classify raw text — 1 credit
curl -X POST https://api.classivore.com/v1/classify \
-H "Authorization: Bearer $CV_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "iab-2.2",
"text": "Tesla announced a new Model Y today...",
"options": { "max_labels": 3, "min_confidence": 0.5 }
}'
Why choose Classivore
Klazify and TextRazor have been around longer. Here's what's different: we take raw text, we batch, and our taxonomies are swappable.
| Feature | Classivore | Klazify | TextRazor | Hive |
|---|---|---|---|---|
| Raw text input | ||||
| URL input (scrape + classify) | partial | |||
| IAB 2.2 taxonomy | v1 only | proprietary | ||
| Custom taxonomies | on request | enterprise | ||
| Batch (100 per request) | ||||
| Predictable per-request pricing | ||||
| No per-token billing |
Rate cards, not sales calls
Every tier is a published number. Caps are soft — you'll get a 429 before you get a surprise bill.
- IAB 2.2 classification
- Playground access
- Community support
- No credit card
- Everything in Free
- Batch endpoint
- Email support
- 90-day usage history
- Everything in Starter
- 12-month usage history
- Everything in Pro
- Dedicated model deployment
- Private region
- Priority support
All tiers soft-block at the cap (HTTP 429). No overage billing, no surprise invoice. Annual billing is 20% off the monthly rate.
Docs that read like a spec, not a blog
Every endpoint, every parameter, every response shape. No tutorials you have to skip.
Classify text
Classify a single text against a taxonomy model. Returns the top-N categories with confidence scores. Text must be 1–50,000 characters.