LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Python ai-agentsdata-engineeringdocument-aigenerative-aiidpjson-extractionllmmcp-serverocrpdf-extractionprompt-engineeringstructured-output
{
"title": "Document classifier tool settings",
"description": "Classify documents based on their content",
"type": "object",
"required": [
"classificationBins"
],
"properties": {
"classificationBins": {
"type": "array",
"title": "Classification bins",
"description": "Specify at least two unique classification bins. `unknown` and `__unstract_failed` are reserved bins.\n - `unknown` indicates the LLM can't determine the classification\n - `__unstract_failed` indicates a tool run failure for the given file.",
"items": {
"type": "string"
},
"minItems": 2,
"uniqueItems": true
},
"useCache": {
"type": "boolean",
"title": "Cache and use cached results",
"default": true,
"description": "Use cached results"
}
}
}