Generate a reusable JSON Schema tailored to a specific document type, allowing for consistent extraction of custom fields across future documents of the same type.
The agent takes a representative text sample of a unique document type and sends it to the schema inference endpoint. It reviews the proposed fields and schema structure, then saves this schema for use in subsequent parsing tasks to ensure consistent output.
- Category: Data and analysis
- Effect: Read-only
- MCP servers: AIDataParser MCP (https://koskamo.com/explore/aidataparser-mcp)
- Source: Docs — AIDataParser API (https://aidataparser.com/docs)
- Source quote: "Infer once, review it, then pass it as schema on parse a document or parse text across every document of that type for consistent output."
- Playbook page: https://koskamo.com/explore/use-cases/infer-custom-schema-from-document-sample
Prompt: Infer a JSON schema from this lab report sample: 'LAB REPORT\nPatient: Jane Doe\nGlucose: 92 mg/dL'. Return the schema and fields.