Excel files support all the same configuration options as regular documents, but some behave differently due to Excel’s native spreadsheet structure.
Quick Summary: Most configuration options work identically to other file types. OCR, high resolution, and pipeline settings are ignored since Excel files use native processing.

​
Configuration Options Overview

Excel configuration options fall into two categories:
CategoryOptionsBehavior
Work NormallySegmentation, Segment Processing, Chunking, LLM Processing, Error Handling, ExpirationSame as other file types with minor Excel-specific notes
IgnoredOCR Strategy, High Resolution, Pipeline ProviderNo effect on Excel processing

​
Options That Work Normally

These configuration options work the same as other file types, with some Excel-specific behavior noted below.

​
Segmentation Strategy

Controls how Excel sheets are analyzed and segmented. 
from chunkr_ai import Chunkr
from chunkr_ai.models import Configuration, SegmentationStrategy

config = Configuration(
    segmentation_strategy=SegmentationStrategy.LAYOUT_ANALYSIS
)

task = chunkr.upload("spreadsheet.xlsx", config)
Available Options:
  • LayoutAnalysis (Recommended): Runs Excel layout analysis to identify tables, charts, and text regions
  • Page: Outputs each full Excel sheet as a single Table segment

​
Segment Processing

Configure how different segment types are processed and formatted. 
from chunkr_ai.models import (
    Configuration,
    SegmentProcessing,
    GenerationConfig,
    SegmentFormat
)

config = Configuration(
    segment_processing=SegmentProcessing(
        Table=GenerationConfig(
            format=SegmentFormat.HTML,
            llm="Extract key insights from this Excel table"
        )
    )
)
Excel-Specific Behavior:
  • Tables: The strategy field (Auto/LLM) is ignored - tables are always extracted natively from Excel
  • LLM Prompts: You can still use custom llm prompts for table processing
  • All Other Segments: Picture, Text, Title, etc. work exactly as with other file types

​
Chunk Processing

Controls how content is divided into chunks for RAG applications. 
from chunkr_ai.models import Configuration, ChunkProcessing

config = Configuration(
    chunk_processing=ChunkProcessing(
        target_chunk_length=1000
    )
)
Excel-Specific Behavior:
  • Works the same as other file types
  • Important: Chunks will break on new sheets (unlike PDFs that chunk across pages)
  • Each Excel worksheet is treated as a boundary for chunking

​
LLM Processing

Configure custom models and prompts for content generation. 
from chunkr_ai.models import Configuration, LlmProcessing

config = Configuration(
    llm_processing=LlmProcessing(
        # Custom LLM configuration
    )
)
Excel Behavior:
  • Works exactly the same as other file types
  • Affects segment processing only
  • Can be combined with segment-specific LLM prompts

​
Error Handling Strategy

Controls how processing errors are handled. 
from chunkr_ai.models import Configuration, ErrorHandlingStrategy

config = Configuration(
    error_handling=ErrorHandlingStrategy.CONTINUE
)
Available Options:
  • Fail: Stop processing on any error
  • Continue: Continue processing despite non-critical errors

​
Expiration Time

Sets how long task results are retained before deletion. 
config = Configuration(
    expires_in=3600  # 1 hour
)
Excel Behavior: Works exactly the same as other file types.

​
Options That Are Ignored

These configuration options have no effect when processing Excel files because Excel uses native processing methods.

​
OCR Strategy

Ignored for Excel files - Excel files contain native text data, so OCR is never applied regardless of this setting.
All OCR-related configurations (All, Auto) are ignored since Excel files provide native text extraction.

​
High Resolution Processing

Ignored for Excel files - Excel files use native resolution processing regardless of this setting.
The high_resolution setting has no effect since Excel files don’t require image-based processing for text content.

​
Pipeline Provider (Azure Feature)

Ignored for Excel files - Excel files always use Chunkr’s native processing pipeline.
Azure Document Intelligence and other pipeline providers are not used for Excel processing.

​
Complete Configuration Example

Here’s a comprehensive configuration example optimized for Excel processing: 
from chunkr_ai import Chunkr
from chunkr_ai.models import (
    Configuration,
    SegmentationStrategy,
    ErrorHandlingStrategy,
    SegmentProcessing,
    GenerationConfig,
    SegmentFormat
)

chunkr = Chunkr()

# Optimal Excel configuration
config = Configuration(
    # Core settings
    segmentation_strategy=SegmentationStrategy.LAYOUT_ANALYSIS,
    error_handling=ErrorHandlingStrategy.CONTINUE,
    
    # Segment processing
    segment_processing=SegmentProcessing(
        Table=GenerationConfig(
            format=SegmentFormat.HTML,
            llm="Analyze this Excel table and extract key insights"
        ),
        Picture=GenerationConfig(
            llm="Describe this Excel chart with key data points"
        )
    ),
    
    # Task settings
    expires_in=7200  # 2 hours
)

task = chunkr.upload("financial_report.xlsx", config)

Understanding Output

Learn about Excel-specific response fields and data structures

Segment Processing

Detailed segment processing configuration options

Chunking

Advanced chunking configuration for RAG applications

LLM Processing

Custom LLM models and prompt configuration