PaperFlow

https://www.paperflowing.com
Admin
PaperFlow is an open-source PDF-to-Markdown engine that converts raw parser output into structured...
Tokens:1,629
Snippets:21
Trust Score:2.1
Update:5 months ago
Show doc for...
Tokens:
Raw
### PaperFlow Local API Quick Start (Bash)

Source: https://www.paperflowing.com/

Commands to clone the PaperFlow repository, install dependencies, and run the local API server.

```bash
git clone https://github.com/TylerMorrison21/paperflow
cd paperflow
pip install -r requirements.txt
uvicorn api.main:app --port 8000
# Open http://localhost:8000
```

--------------------------------

### Raw Parser Output Example

Source: https://www.paperflowing.com/

An example of raw text output from a PDF parser, highlighting issues like unclickable references and incorrect delimiters.

```text
[1] dead text -- can't click
\[ E = mc^2 \] wrong delimiters
[Fig. 3] plain string
No metadata

References
[1] Vaswani et al. Attention Is All You Need.
```

--------------------------------

### PaperFlow Processed Markdown Example

Source: https://www.paperflowing.com/

The same content as the raw output, but processed by PaperFlow to include features like hoverable references, renderable LaTeX, and clickable figure links.

```markdown
[^1] hover to preview reference
$$ E = mc^2 $$ renders everywhere
[[#^fig-3|Fig. 3]] click to jump
---
title: "Attention Is All You Need"
authors:
  - "Vaswani"
  - "Shazeer"
---
```

--------------------------------

### PaperFlow Local Web UI Workflow Diagram

Source: https://www.paperflowing.com/

A visual representation of the PaperFlow Local Web UI, guiding users through parser selection, PDF submission, and download.

```text
+---------------------------------------------+
|  PaperFlow Local UI                         |
|                                             |
|  Quick decision guide                       |
|  Standard digital PDF? -> PyMuPDF           |
|  Scan / formulas / OCR? -> PaddleOCR-VL     |
|  Easiest premium path? -> Marker API        |
|  Private enterprise setup? -> Self-hosted   |
|                                             |
|  1. Choose parser                           |
|     (*) PyMuPDF Local                       |
|     ( ) PaddleOCR-VL-0.9B                   |
|     ( ) Marker API (Datalab.to)             |
|     ( ) Enterprise Marker Self-Hosted       |
|                                             |
|  2. Drop PDF here                           |
|     +-------------------------------+       |
|     |  paper.pdf  /  2.4 MB         |       |
|     +-------------------------------+       |
|                                             |
|  Setup, tradeoffs, and exact commands       |
|  appear for the selected parser             |
|                                             |
|  [Convert]  [Batch Process]                 |
|                                             |
|  3. Download                                |
|      paper.md     paperflow.zip             |
+---------------------------------------------+
```

--------------------------------

### PaperFlow MCP Server Configuration (Windows)

Source: https://www.paperflowing.com/

JSON configuration for setting up PaperFlow as an MCP server on Windows systems.

```json
{
  "mcpServers": {
    "paperflow": {
      "command": "cmd",
      "args": ["/c", "npx", "-y", "paperflow-mcp"]
    }
  }
}
```

--------------------------------

### PaperFlow MCP Server Configuration (macOS/Linux)

Source: https://www.paperflowing.com/

JSON configuration for setting up PaperFlow as an MCP server on macOS or Linux systems.

```json
{
  "mcpServers": {
    "paperflow": {
      "command": "npx",
      "args": ["-y", "paperflow-mcp"]
    }
  }
}
```

--------------------------------

### PaperFlow Core Idea Workflow

Source: https://www.paperflowing.com/

Illustrates the conceptual flow of processing raw parser output into structured Markdown using PaperFlow.

```text
raw parser output
  -> normalize
  -> enrich
  -> structured markdown
```

--------------------------------

### Submitting PDF to PaperFlow API (Bash)

Source: https://www.paperflowing.com/

Bash commands to submit a PDF file to the local PaperFlow API and retrieve the processed Markdown and package.

```bash
curl -X POST http://localhost:8000/api/submit \
  -F "file=@paper.pdf"

curl http://localhost:8000/api/jobs/<job_id>/result -o paper.md
curl http://localhost:8000/api/jobs/<job_id>/package -o paperflow.zip
```

--------------------------------

### PaperFlow Post-processing Python API

Source: https://www.paperflowing.com/

Python code demonstrating how to use the `paperflow-postprocess` package to enhance raw Markdown output with metadata.

```python
from paperflow_postprocess import enhance

raw_md = open("parser_output.md").read()
result = enhance(raw_md, images={}, metadata={"title": "My Paper"})
```
Tokens:
Raw