Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.runpulse.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

Meridian is Pulse’s built-in table-to-Excel conversion engine. After extracting a document, any tables detected by Pulse can be exported as a formatted .xlsx workbook — ready for downstream analysis, reporting, or integration.

​
When to Use

  • Financial data — export income statements, balance sheets, or transaction tables to Excel
  • Invoices — convert line item tables to spreadsheets for accounting systems
  • Research data — pull experimental results or survey tables into a workable format
  • Any document with tables — if Pulse detects tables in the extraction, you can export them

​
How to Use in the Playground

1
Extract a document
2
Upload and extract your document using any pipeline — Extract, Extract → Schema, or Extract → Split → Schema.
4
After extraction, switch to the Tables tab to see all detected tables rendered in a grid view.
5
Click “Export to Excel”
6
Click the Export button (spreadsheet icon) in the extraction results toolbar. This opens the Meridian export flow.
7
Select tables and configure
8
In the export view, you can:
9
  • Select/deselect tables — choose which tables to include
  • Rename sheets — customize the Excel sheet name for each table
  • Choose export mode — separate workbooks per document, combined into one, or appended to an existing workbook
    • 10
    Configure formatting
    11
    Set Excel formatting preferences:
    12
    SettingOptionsFont familyArial, Calibri, Times New Roman, etc.Font size8–16ptFont colorAny hex colorAlignmentLeft, Center, RightBold / Italic / UnderlineToggle eachColumn widthAuto or fixedRow heightAuto or fixed
    13
    Export
    14
    Click Export All. The workbook is generated and downloaded to your browser as a .xlsx file.

    ​
    Export Modes

    ModeDescription
    SeparateOne .xlsx file per document. Each table gets its own sheet within the workbook.
    CombinedAll tables from all documents merged into a single .xlsx file. Each table gets its own sheet.
    AppendedTables are added to an existing .xlsx workbook you upload. You can append to an existing sheet (vertically or horizontally) or create new sheets.

    ​
    Appending to Existing Workbooks

    The Appended mode is uniquely powerful — upload an existing Excel file and Meridian will add the newly extracted tables to it:
    • Vertical append — new rows added below existing data in a target sheet
    • Horizontal append — new columns added to the right of existing data
    • New tab — each table becomes a new sheet in the existing workbook
    Options for vertical append:
    • Include unmatched columns — add columns from the new table that don’t exist in the target
    • Add identifier column — insert a column to tag rows by their source document

    ​
    API Usage

    The Meridian conversion is available via the /meridian-convert endpoint: 
    import requests

# After extraction, you have bounding_boxes with table data
tables = [
    {
        "name": "Income Statement",
        "rows": 5,
        "columns": 3,
        "page": 1,
        "cells": [
            {"rowIndex": 0, "columnIndex": 0, "content": "Category"},
            {"rowIndex": 0, "columnIndex": 1, "content": "Q1"},
            {"rowIndex": 0, "columnIndex": 2, "content": "Q2"},
            {"rowIndex": 1, "columnIndex": 0, "content": "Revenue"},
            {"rowIndex": 1, "columnIndex": 1, "content": "1200000"},
            {"rowIndex": 1, "columnIndex": 2, "content": "1350000"}
        ],
        "customSheetName": "Income Statement"
    }
]

response = requests.post(
    "https://api.runpulse.com/meridian-convert",
    headers={"x-api-key": "YOUR_API_KEY"},
    json={
        "tables": tables,
        "filename": "financial_tables.xlsx",
        "formatSettings": {
            "fontFamily": "Calibri",
            "fontSize": 11,
            "alignment": "left"
        }
    }
)

# Response contains a presigned download URL
download_url = response.json()["presigned_url"]
print(f"Download: {download_url}")

    ​
    Response

    {
  "s3_object_url": "https://bucket.s3.amazonaws.com/meridian-convert/.../financial_tables.xlsx",
  "presigned_url": "https://bucket.s3.amazonaws.com/meridian-convert/...?X-Amz-Signature=...",
  "filename": "financial_tables.xlsx",
  "content_type": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
}
    The presigned_url is valid for 7 days and can be used to download the file directly.

    ​
    End-to-End: Extract → Meridian

    A common workflow is to extract a document with tables and immediately convert them to Excel: 
    from pulse import Pulse
import requests

client = Pulse(api_key="YOUR_API_KEY")

# 1. Extract the document
result = client.extract(
    file=open("financial_report.pdf", "rb"),
    storage={"enabled": True}
)

# 2. Get tables from bounding_boxes
bounding_boxes = result.bounding_boxes or {}
raw_tables = bounding_boxes.get("Tables", [])

# 3. Convert to Meridian table format
tables = []
for i, tbl in enumerate(raw_tables):
    cells = [
        {
            "rowIndex": c["position"]["row"],
            "columnIndex": c["position"]["column"],
            "content": c.get("text", "")
        }
        for c in tbl.get("cell_data", [])
    ]
    tables.append({
        "name": f"Table {i+1}",
        "rows": tbl.get("table_info", {}).get("dimensions", {}).get("rows", 0),
        "columns": tbl.get("table_info", {}).get("dimensions", {}).get("columns", 0),
        "page": 1,
        "cells": cells,
        "customSheetName": f"Table {i+1}"
    })

# 4. Export to Excel
resp = requests.post(
    "https://api.runpulse.com/meridian-convert",
    headers={"x-api-key": "YOUR_API_KEY"},
    json={"tables": tables, "filename": "report_tables.xlsx"}
)

print(f"Download: {resp.json()['presigned_url']}")

    Extract Pipeline

    The extraction step that produces the tables for Meridian

    Bounding Boxes

    Understanding the bounding box output that powers table detection