Skip to content

Getting Started

The Split API provides a simple way to automatically segment concatenated PDFs into logical document sections based on content categories. Using AI-powered classification, it analyzes each page’s content, classifies pages into user-defined categories, and groups consecutive pages of the same category into segments.

Split is currently in beta and is subject to breaking changes. SDK support is not yet available—all interactions use the REST API directly.

Split is a great fit when you need:

  • Document separation: Automatically separate bundled documents (e.g., a collection of reports, research papers, or contracts) before further processing
  • Content organization: Categorize and organize mixed document collections by type
  • Pre-processing for extraction: Identify different document types within a single file before running targeted extraction with LlamaExtract
  • Flexible categorization: Define any categories relevant to your use case with natural-language descriptions
  • AI-powered classification: Uses LLMs to understand page content and assign categories
  • Flexible categories: Define any categories relevant to your use case
  • Confidence scoring: Each segment includes a confidence level (high, medium, low)
  • Page-level granularity: Results include exact page numbers for each segment
  • Uncategorized handling: Optionally capture pages that don’t match any defined category

First, get an API key to use the Split API.

Since SDK support is not yet available, you’ll interact with Split using the REST API directly. Check out the REST API guide for a complete walkthrough.

  1. Upload your PDF to LlamaCloud using the Files API
  2. Define categories with names and natural-language descriptions
  3. Create a split job with your file and categories
  4. Poll for completion until the job finishes
  5. Retrieve results with segments, page ranges, and confidence scores

See Pricing for credit rates and billing details.

  • Create Split Job: POST /api/v1/beta/split/jobs
  • Get Split Job: GET /api/v1/beta/split/jobs/{job_id}
  • List Split Jobs: GET /api/v1/beta/split/jobs

For the complete API reference, see our API documentation.