Parseflow Process Request

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://docs.parseflow.tech/schemas/process-request.json",
  "title": "Parseflow Process Request",
  "description": "Multipart form payload accepted by POST /v2/process. Either `file` or `text` must be supplied.",
  "type": "object",
  "properties": {
    "file": {
      "type": "string",
      "contentMediaType": "application/octet-stream",
      "description": "Binary document upload (PDF, DOCX, or TXT)."
    },
    "text": {
      "type": "string",
      "description": "Raw text payload used in lieu of a file upload."
    },
    "chunk_size": {
      "type": "integer",
      "minimum": 200,
      "maximum": 10000,
      "default": 2000,
      "description": "Maximum chunk length in characters."
    },
    "overlap": {
      "type": "integer",
      "minimum": 0,
      "maximum": 2000,
      "default": 200,
      "description": "Character overlap between adjacent chunks."
    },
    "mode": {
      "type": "string",
      "enum": ["deterministic", "byok_assisted", "basic"],
      "default": "deterministic"
    },
    "output_format": {
      "type": "string",
      "enum": ["json", "markdown", "zip"],
      "default": "json"
    },
    "include_markdown": {
      "type": "boolean",
      "default": false
    },
    "preset": {
      "type": ["string", "null"],
      "description": "Named extraction preset, e.g. invoice, receipt, contract."
    },
    "schema_json": {
      "type": ["string", "null"],
      "description": "JSON-encoded schema for caller-driven structured extraction."
    },
    "enforce_schema": {
      "type": "boolean",
      "default": false
    },
    "byok_provider": {
      "type": ["string", "null"],
      "enum": ["openai", "anthropic", "google", null]
    },
    "byok_model": {
      "type": ["string", "null"]
    }
  },
  "anyOf": [
    { "required": ["file"] },
    { "required": ["text"] }
  ]
}