@uploadista/flow-documents-nodes

Document processing nodes for Uploadista flows. Process PDFs with text extraction, OCR, splitting, merging, and conversion.

Installation

npm install @uploadista/flow-documents-nodes
# or
pnpm add @uploadista/flow-documents-nodes

Quick Start

import {
  createOcrNode,
  createExtractTextNode,
  createSplitPdfNode,
  createMergePdfNode,
  createDescribeDocumentNode,
  createConvertToMarkdownNode,
} from "@uploadista/flow-documents-nodes";

Node Types

OCR Node

AI-powered text extraction from scanned documents.

import { createOcrNode } from "@uploadista/flow-documents-nodes";

// Convert scanned document to markdown
const ocrNode = yield* createOcrNode("ocr-1", {
  taskType: "convertToMarkdown",
  resolution: "gundam",
  credentialId: "my-replicate-credential",
});

// Free-form OCR for plain text extraction
const freeOcrNode = yield* createOcrNode("ocr-2", {
  taskType: "freeOcr",
  resolution: "base",
});

// Locate specific content in document
const locateNode = yield* createOcrNode("ocr-3", {
  taskType: "locateObject",
  referenceText: "Invoice Total",
  resolution: "small",
});

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | taskType | "convertToMarkdown" \| "freeOcr" \| "parseFigure" \| "locateObject" | Yes | - | OCR task type | | resolution | "tiny" \| "small" \| "base" \| "gundam" \| "large" | No | - | Model resolution | | credentialId | string | No | - | AI service credential ID | | referenceText | string | No | - | Text to locate (for locateObject task) | | keepOutput | boolean | No | false | Keep output in flow results |

Task Types

| Task Type | Description | |-----------|-------------| | convertToMarkdown | Structured markdown output with headings, lists | | freeOcr | Unstructured plain text extraction | | parseFigure | Analyze charts and diagrams | | locateObject | Find specific content using reference text |

Extract Text Node

Fast text extraction from searchable PDFs.

import { createExtractTextNode } from "@uploadista/flow-documents-nodes";

// Extract text from searchable PDF
const extractNode = yield* createExtractTextNode("extract-1");

// With keepOutput option
const keepOutputNode = yield* createExtractTextNode("extract-2", {
  keepOutput: true,
});

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | keepOutput | boolean | No | false | Keep output in flow results |

Output: Adds extractedText to file metadata.

Split PDF Node

Split PDFs by page range or into individual pages.

import { createSplitPdfNode } from "@uploadista/flow-documents-nodes";

// Extract pages 3-5 as single PDF
const rangeNode = yield* createSplitPdfNode("split-1", {
  mode: "range",
  startPage: 3,
  endPage: 5,
});

// Split each page into separate PDF
const individualNode = yield* createSplitPdfNode("split-2", {
  mode: "individual",
});

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | mode | "range" \| "individual" | Yes | - | Split mode | | startPage | number | No | - | Start page (for range mode) | | endPage | number | No | - | End page (for range mode) | | keepOutput | boolean | No | false | Keep output in flow results | | naming | FileNamingConfig | No | - | File naming configuration |

Merge PDF Node

Combine multiple PDFs into a single document.

import { createMergePdfNode } from "@uploadista/flow-documents-nodes";

// Merge PDFs with default settings
const mergeNode = yield* createMergePdfNode("merge-1");

// With custom naming
const namedMergeNode = yield* createMergePdfNode("merge-2", {
  naming: { mode: "auto" },
});

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | inputCount | number | No | - | Expected number of input files | | keepOutput | boolean | No | false | Keep output in flow results | | naming | FileNamingConfig | No | - | File naming (auto suffix: merged) |

Note: Requires a Merge utility node upstream to provide multiple files.

Describe Document Node

Extract PDF metadata (page count, author, title, etc.).

import { createDescribeDocumentNode } from "@uploadista/flow-documents-nodes";

// Extract document metadata
const describeNode = yield* createDescribeDocumentNode("describe-1");

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | keepOutput | boolean | No | false | Keep output in flow results |

Output Metadata:

{
  "pageCount": 10,
  "format": "pdf",
  "author": "John Doe",
  "title": "Document Title",
  "subject": "Document Subject",
  "creator": "Adobe Acrobat",
  "creationDate": "2023-01-01T00:00:00Z",
  "modifiedDate": "2023-01-02T00:00:00Z",
  "fileSize": 1024000
}

Convert to Markdown Node

Intelligent document-to-markdown conversion.

import { createConvertToMarkdownNode } from "@uploadista/flow-documents-nodes";

// Convert with default settings
const convertNode = yield* createConvertToMarkdownNode("convert-1");

// With custom resolution and credential
const customNode = yield* createConvertToMarkdownNode("convert-2", {
  resolution: "gundam",
  credentialId: "my-ai-credential",
});

Parameters

| Parameter | Type | Required | Default | Description | |-----------|------|----------|---------|-------------| | resolution | "tiny" \| "small" \| "base" \| "gundam" \| "large" | No | "gundam" | OCR model resolution | | credentialId | string | No | - | AI service credential ID | | keepOutput | boolean | No | false | Keep output in flow results |

How it Works:

1. Tries text extraction first (fast, for searchable PDFs)
2. Falls back to OCR if no text found (for scanned PDFs)
3. Returns structured markdown in metadata.markdown

Requirements

• DocumentPlugin: Required for PDF operations (split, merge, extract text, metadata)
• DocumentAiPlugin: Required for OCR and AI-powered conversion

License

MIT