The Split API intelligently divides documents into sections based on content categories, making it easy to organize and process different parts of a document separately.
Basic Usage
from reducto import Reducto
client = Reducto()
response = client.split.run(
input="https://example.com/document.pdf",
split_description=[
{"name": "summary", "description": "Executive summary section"},
{"name": "financials", "description": "Financial data and tables"},
{"name": "risks", "description": "Risk factors and disclosures"}
]
)
print(response)
import asyncio
from reducto import AsyncReducto
client = AsyncReducto()
async def main():
response = await client.split.run(
input="https://example.com/document.pdf",
split_description=[
{"name": "summary", "description": "Executive summary section"},
{"name": "financials", "description": "Financial data and tables"},
{"name": "risks", "description": "Risk factors and disclosures"}
]
)
print(response)
asyncio.run(main())
Method Signature
client.split.run(
input: str,
split_description: list[SplitCategory],
parsing: ParseOptions | None = None,
settings: dict | None = None,
split_rules: str | None = None
) -> SplitResponse
Parameters
The URL of the document to split. You can provide:
- A publicly available URL
- A presigned S3 URL
- A
reducto:// prefixed URL from the /upload endpoint
- A
jobid:// prefixed URL from a previous parse invocation
- A list of URLs (for multi-document pipelines, V3 API only)
List of category definitions for splitting the document. Each category should have:
name: Category identifierdescription: Description of what content belongs in this category
Configuration options for parsing the document. If you’re passing in a jobid:// URL, this will be ignored.
Settings for split processing.
Natural language prompt describing rules for splitting the document.
Split Categories
Define categories based on document structure:
from reducto import Reducto
client = Reducto()
categories = [
{
"name": "introduction",
"description": "Opening remarks and overview"
},
{
"name": "methodology",
"description": "Research methods and approach"
},
{
"name": "results",
"description": "Findings, data, and analysis"
},
{
"name": "conclusion",
"description": "Summary and final thoughts"
}
]
response = client.split.run(
input="https://example.com/research-paper.pdf",
split_description=categories
)
# Access split sections
for section in response.sections:
print(f"{section.category}: {len(section.content)} chars")
import asyncio
from reducto import AsyncReducto
client = AsyncReducto()
async def main():
categories = [
{
"name": "introduction",
"description": "Opening remarks and overview"
},
{
"name": "methodology",
"description": "Research methods and approach"
},
{
"name": "results",
"description": "Findings, data, and analysis"
},
{
"name": "conclusion",
"description": "Summary and final thoughts"
}
]
response = await client.split.run(
input="https://example.com/research-paper.pdf",
split_description=categories
)
# Access split sections
for section in response.sections:
print(f"{section.category}: {len(section.content)} chars")
asyncio.run(main())
Custom Split Rules
Add natural language rules to guide the splitting process:
from reducto import Reducto
client = Reducto()
response = client.split.run(
input="https://example.com/contract.pdf",
split_description=[
{"name": "terms", "description": "Terms and conditions"},
{"name": "pricing", "description": "Pricing and payment terms"},
{"name": "warranties", "description": "Warranties and guarantees"}
],
split_rules="Split at major section boundaries. Keep related clauses together. Preserve the hierarchy of subsections."
)
Split with Parsing Options
Combine splitting with custom parsing configuration:
from reducto import Reducto
client = Reducto()
response = client.split.run(
input="https://example.com/document.pdf",
split_description=[
{"name": "text_sections", "description": "Text-heavy sections"},
{"name": "data_sections", "description": "Sections with tables and charts"}
],
parsing={
"enhance": {
"summarize_figures": True,
"agentic": ["table"]
},
"formatting": {
"table_output_format": "json",
"add_page_markers": True
}
}
)
Async Job Processing
For large documents, use async job processing:
from reducto import Reducto
client = Reducto()
# Start an async split job
job = client.split.run_job(
input="https://example.com/large-document.pdf",
split_description=[
{"name": "section1", "description": "First section"},
{"name": "section2", "description": "Second section"}
],
async_={
"webhook": {"url": "https://example.com/webhook"}
}
)
print(f"Job ID: {job.job_id}")
# Poll for results
result = client.job.get(job.job_id)
Financial Document Example
Split a financial report into meaningful sections:
from reducto import Reducto
client = Reducto()
categories = [
{
"name": "executive_summary",
"description": "High-level overview and key highlights"
},
{
"name": "financial_statements",
"description": "Income statement, balance sheet, cash flow"
},
{
"name": "md_and_a",
"description": "Management discussion and analysis"
},
{
"name": "footnotes",
"description": "Accounting notes and disclosures"
},
{
"name": "risk_factors",
"description": "Risk disclosures and forward-looking statements"
}
]
response = client.split.run(
input="https://example.com/annual-report.pdf",
split_description=categories,
split_rules="Preserve table integrity. Keep footnote references with their corresponding sections."
)
# Process each section separately
for section in response.sections:
print(f"\n=== {section.category.upper()} ===")
print(section.content[:200] + "...")
Reusing Parsed Documents
Split a document that was previously parsed:
from reducto import Reducto
client = Reducto()
# First parse the document
parse_response = client.parse.run(
input="https://example.com/document.pdf",
formatting={"add_page_markers": True}
)
# Then split using the job ID (no re-parsing needed)
split_response = client.split.run(
input=f"jobid://{parse_response.job_id}",
split_description=[
{"name": "part1", "description": "First part"},
{"name": "part2", "description": "Second part"}
]
)