Biblicus

Make your documents usable by your assistant, then decide later how you will search and retrieve them.

SmolDocling-256M Extractor

Extractor ID: docling-smol

Category: Vision-Language Models (VLM)

Overview

The SmolDocling-256M extractor uses IBM Research's SmolDocling vision-language model for fast, accurate document understanding. It combines visual layout analysis with semantic text extraction to handle complex documents.

SmolDocling-256M is a 256-million parameter VLM optimized for document processing. It achieves 6.15 seconds per page on Apple Silicon with MLX, making it one of the fastest VLM extractors while maintaining excellent accuracy.

Installation

Transformers Backend (Cross-Platform)

pip install "biblicus[docling]"
pip install "biblicus[docling-mlx]"

The MLX backend provides 2-3x faster inference on Apple Silicon (M1/M2/M3/M4) with lower memory usage.

Supported Media Types

The extractor automatically skips text items (text/plain, text/markdown) and audio items.

Configuration

Config Schema

class DoclingSmolExtractorConfig(BaseModel):
    output_format: str = "markdown"  # markdown, text, or html
    backend: str = "mlx"              # mlx or transformers

Configuration Options

Option Type Default Description
output_format str markdown Output format: markdown, text, or html
backend str mlx Inference backend: mlx (Apple Silicon) or transformers (cross-platform)

Output Formats

Usage

Command Line

Basic Usage

# Extract using SmolDocling with defaults (markdown, MLX)
biblicus extract my-corpus --extractor docling-smol

Custom Configuration

# Use Transformers backend with HTML output
biblicus extract my-corpus --extractor docling-smol \
  --config output_format=html \
  --config backend=transformers

Configuration File

extractor_id: docling-smol
config:
  output_format: markdown
  backend: mlx
biblicus extract my-corpus --configuration configuration.yml

Python API

from biblicus import Corpus

# Load corpus
corpus = Corpus.from_directory("my-corpus")

# Extract with defaults
results = corpus.extract_text(extractor_id="docling-smol")

# Extract with custom config
results = corpus.extract_text(
    extractor_id="docling-smol",
    config={
        "output_format": "html",
        "backend": "transformers"
    }
)

In Pipeline

Fallback Chain

extractor_id: select-text
config:
  extractors:
    - pdf-text         # Try text extraction first
    - docling-smol     # Fall back to VLM

Media Type Routing

extractor_id: select-smart-override
config:
  default_extractor: pdf-text
  overrides:
    - media_type_pattern: "image/.*"
      extractor: docling-smol

Examples

Academic Papers

Extract academic papers with equations and code blocks:

biblicus extract papers-corpus --extractor docling-smol \
  --config output_format=markdown

Office Documents

Process DOCX, XLSX, PPTX files:

biblicus extract office-corpus --extractor docling-smol \
  --config output_format=html

Scanned Documents

OCR scanned PDFs and images:

biblicus extract scans-corpus --extractor docling-smol \
  --config backend=mlx

Multi-Format Corpus

Handle mixed document types with automatic routing:

from biblicus import Corpus

corpus = Corpus.from_directory("mixed-corpus")

# Route based on media type
results = corpus.extract_text(
    extractor_id="select-smart-override",
    config={
        "default_extractor": "pass-through-text",
        "overrides": [
            {"media_type_pattern": "application/pdf", "extractor": "docling-smol"},
            {"media_type_pattern": "image/.*", "extractor": "docling-smol"},
            {"media_type_pattern": "application/vnd\\.openxmlformats.*", "extractor": "docling-smol"},
        ]
    }
)

Performance

Benchmarks

Backend Comparison

Backend Platform Speed Memory
MLX Apple Silicon 6.15 sec/page Efficient
Transformers Any (CPU/CUDA) 15-20 sec/page Higher

When to Use SmolDocling vs Granite

Error Handling

Missing Dependency

If the Docling library is not installed:

ExtractionRunFatalError: DoclingSmol extractor requires an optional dependency.
Install it with pip install "biblicus[docling]".

Missing MLX Support

If MLX backend is configured but not available:

ExtractionRunFatalError: DoclingSmol extractor with MLX backend requires MLX support.
Install it with pip install "biblicus[docling-mlx]".

Empty Output

Documents that cannot be processed produce empty extracted text and are counted in extracted_empty_items statistics.

Per-Item Errors

Processing errors for individual items are recorded in the extraction snapshot but don't halt the entire extraction. Check errored_items in extraction statistics.

Same Category

Alternatives

Pipeline Utilities

See Also