Skip to main content

​
Overview

The Extraction Library is where all your saved extractions live. Every time you extract a document with storage: { enabled: true }, the result is saved to your organization’s library. You can browse past extractions, organize them into folders, re-open results, run schema or split on them, and share them with team members.

​
Accessing the Library

Navigate to Extractions in the Playground sidebar to open the library. You’ll see:
  • Folders — organize extractions into named folders
  • Extractions — individual saved extractions with filename, file type, page count, and creation date
  • Schema count — number of schema versions applied to each extraction

​
Features

​
Search and Filter

Use the search bar at the top to find extractions by filename. Results filter in real-time across all folders.

​
Folders

Organize your extractions into folders for different projects, clients, or document types:
  • Create folders — click the New Folder button
  • Drag and drop — move extractions between folders by dragging
  • Expand/collapse — click a folder to see its contents
  • Rename/delete — right-click or use the menu on each folder

​
Pagination

Large libraries are paginated automatically:
  • Folders load in batches of 10, with a “Load More” button
  • Extractions within folders load in batches of 50, with pagination controls

​
Viewing an Extraction

Click any extraction to open it in the detail view. From there you can:
  • View markdown — see the full extracted text
  • View tables — browse detected tables
  • View bounding boxes — see element locations overlaid on the original document
  • View schema — see structured output (if schema was applied)
  • Run new schema — apply a new or modified schema without re-extracting
  • Run split — divide the extraction into topics
  • Re-extract — process the document again with different settings
  • Export to Excel — convert tables to .xlsx via Meridian

​
Sharing

Share an extraction with a public link:
  • Click the Share button on any extraction
  • Toggle public access on/off
  • Set an optional expiration date
  • Copy the public URL to share with anyone — no API key required

​
Bulk Actions

Select multiple extractions using checkboxes for bulk operations:
  • Move to folder — move selected extractions to a different folder
  • Delete — remove selected extractions

​
Schema Versioning

Each extraction can have multiple schema versions. When you apply a new schema (or re-run with a modified one), a new version is created — the previous version is preserved. In the extraction detail view:
  • Switch between versions using the version selector
  • Compare results across different schema configurations
  • Roll back to a previous version if needed
See Reruns for more on iterating with schemas.

​
API: Listing Saved Extractions

You can also list and retrieve saved extractions programmatically: 
import requests

response = requests.get(
    "https://api.runpulse.com/api/v1/extractions",
    headers={"x-api-key": "YOUR_API_KEY"},
    params={"folders_limit": 10, "per_folder_limit": 50}
)

data = response.json()

for folder in data["folders_with_extractions"]:
    print(f"Folder: {folder['name']}")
    for extraction in folder["extractions"]:
        print(f"  - {extraction['filename']} ({extraction['page_count']} pages)")

​
Storage Requirement

Extractions only appear in the library if storage: { enabled: true } was passed during extraction. Without storage, the extraction result is returned directly but not saved.
To ensure all extractions are saved, always include: 
result = client.extract(
    file=open("document.pdf", "rb"),
    storage={"enabled": True}
)
Or in the Playground, storage is enabled by default.

Reruns

Re-run schema, split, or extraction on saved documents

Step Preset Library

Save and reuse pipeline configurations