Extract & structure ↗
Classification
Route or tag text against your own taxonomy — a label, a confidence, and a one-line rationale.
POST /v1/classify$0.02/call · 2 credits, charged only on success
Send this
{
"text": "My card was charged twice for one order.",
"labels": [
"billing",
"shipping",
"technical",
"refund"
]
}Every response carries usage — the credits burned and your remaining balance. A failed call never burns credits.
Try it live against your own key in the interactive reference; every account gets 500 free credits a month, no card.
Get this back
{
"label": "billing",
"confidence": 0.96,
"rationale": "Duplicate charge is a billing dispute.",
"usage": {
"credits": 2,
"balanceRemaining": 484
}
}Four ways in. Same $0.02/call.
curl
curl https://api.kynth.studio/v1/classify \
-H "Authorization: Bearer ksk_live_…" \
-H "Content-Type: application/json" \
-d '{"text":"My card was charged twice for one order.","labels":["billing","shipping","technical","refund"]}'TypeScript
import { KynthCore } from "@kynth/api";
const kynth = new KynthCore({ apiKey: process.env.KYNTH_API_KEY! });
const result = await kynth.classify({
"text": "My card was charged twice for one order.",
"labels": [
"billing",
"shipping",
"technical",
"refund"
]
});Python
from kynth import Kynth
client = Kynth(api_key=os.environ["KYNTH_API_KEY"])
result = client.classify(
text="My card was charged twice for one order.",
labels=[
"billing",
"shipping",
"technical",
"refund"
],
)MCP
// One-time setup in your MCP client (Claude, Cursor, …):
{ "mcpServers": { "kynth-core": {
"command": "npx", "args": ["@kynth/api-mcp"],
"env": { "KYNTH_API_KEY": "ksk_live_…" } } } }
// Then the agent just calls the kynth_classify tool.More from extract & structure
Field extraction
$0.04/call
Pull a field set you define out of any block of text. You name the fields; you get typed values with confidence.
Guaranteed structure
$0.06/call
Any messy input — text, HTML, an email — plus YOUR JSON schema → output that conforms to it. Retries and validation handled.
Record normalization
$0.05/batch
A batch of messy records → clean canonical rows, with a change log of every fix (casing, formats, dedup-ready values).