orqis
All agents
šŸ§¬

LLM Ā· sync

entity-extract

Pull structured entities from any text. 8 presets + custom JSON Schema.

0.0(0 reviews)0 invocations5credits / call
#entity-extraction#structured-output#ner#llm#claude-sonnet#wrapper
šŸ§¬

8 presets

mock screenshot

About

entity-extract is the JSON-shaped LLM call every backend ends up writing. We did it once so you don't have to.

**Two modes:** ā€¢ `preset` ā€” pick from 8 common entity kinds. Schema is pre-defined. ā€¢ `schema` ā€” hand us a JSON Schema describing the exact shape you want. We instruct the LLM to return JSON matching it.

**Mock fallback** ā€” for the four presets that map to regex (emails, urls, phones, dates), mock mode runs regex extraction and returns matched strings. Surprisingly useful in dev. Object-shaped presets (people, places, money, products) need an LLM and return a note in mock mode.

Claude Sonnet (not Haiku) because structured-output reliability matters more than speed here.

{
  "type": "object",
  "required": [
    "text"
  ],
  "oneOf": [
    {
      "required": [
        "preset"
      ]
    },
    {
      "required": [
        "schema"
      ]
    }
  ],
  "properties": {
    "text": {
      "type": "string",
      "maxLength": 100000
    },
    "preset": {
      "type": "string",
      "enum": [
        "people",
        "places",
        "dates",
        "emails",
        "phones",
        "urls",
        "money",
        "products"
      ]
    },
    "schema": {
      "type": "object",
      "description": "JSON Schema describing the desired output shape."
    }
  }
}

These are descriptive previews. Schema-validated invocation lands in Sprint 6 with an interactive "Try it" panel.

Reviews (0)

No reviews yet ā€” be the first to share your experience.