> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.instabase.com/flow/step-config-reference/agent-classifier/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.instabase.com/_mcp/server.

# Agent classifier

> Classify records using a large language model (LLM) and your classification schema.

Enterprise Single-tenant

The *Agent classifier* step assigns each record to a document class using an LLM, according to the schema defined in a linked agent classifier module.

## Parameters

* **Schema configuration** -- Select the agent classifier module containing your classification schema JSON file.

  The JSON shape matches the extraction schema used by the [*Agent extract* step](/flow/step-config-reference/agent-extraction/): top-level keys are class names, and each class value has `description` and `fields`. You can reuse the same field objects as in your *Agent extract* module's JSON schema, or list only the field names the classifier must consider.

  Every class must define `fields` as a non-empty array. Each entry must at least include `name`. Omitting `fields` or leaving it empty can cause classification to fail.

## Sample classification schemas

The following schema defines field names only.

```json
{
  "Invoice": {
    "description": "An invoice document requesting payment for goods or services",
    "fields": [
      { "name": "Invoice Number" }
    ]
  }
}
```

The following schema provides full field objects with the same structure as seen in the *Agent extract* step.

```json
{
  "Invoice": {
    "description": "An invoice document requesting payment for goods or services",
    "fields": [
      {
        "name": "Invoice Number",
        "data_type": "TEXT",
        "description": "The unique invoice identifier or number"
      }
    ]
  }
}
```

> Classify records using a large language model (LLM) and your classification schema.