Validation Failed

Required variable 'pages' is not in the inputs of pipe 'summarize_document'. Current inputs: - document: native.Document

doc_summarizer

Document Summarizer

v0.1.3MITValidation Failed8 installs3/2/2026Pipelex/methodsEvotis S.A.S
Main pipesummarize_document

Install

$npx mthds install https://github.com/Pipelex/methods --method doc_summarizer

Run

$npx mthds run method doc_summarizer

Document Summarizer

Deep document summarization: profile the document, extract key points with importance ranking, then synthesize into a structured summary with themes and open questions.

What it does

Given a document (PDF, Word, etc.), this method:

  1. Extracts content from the document into readable pages
  2. In parallel:
    • Profiles the document structure, type, language, and organization → DocumentProfile
    • Extracts key points with importance ranking (critical/high/medium/low), categorization, and supporting evidence → KeyPoint[]
  3. Synthesizes the profile and key points into a comprehensive Summary with abstract, ranked key points, themes, and open questions

Concepts

  • DocumentProfile: Metadata — title, document_type, structure_type, language, sections_overview
  • KeyPoint: Individual finding with point, importance, category, supporting_evidence
  • Summary: Final synthesis with abstract, key_points[], themes[], open_questions[]

Exports

document_summarization

summarize_documentprofile_documentextract_key_pointssynthesize_summary

Bundle Files

Select a file to view its contents

METHODS.toml

METHODS.toml
[package]
name = "doc_summarizer"
display_name = "Document Summarizer"
address = "github.com/pipelex/methods"
version = "0.1.3"
description = "# Document Summarizer\n\nDeep document summarization: profile the document, extract key points with importance ranking, then synthesize into a structured summary with themes and open questions.\n\n## What it does\n\nGiven a **document** (PDF, Word, etc.), this method:\n\n1. **Extracts** content from the document into readable pages\n2. **In parallel**:\n   - **Profiles** the document structure, type, language, and organization → `DocumentProfile`\n   - **Extracts key points** with importance ranking (critical/high/medium/low), categorization, and supporting evidence → `KeyPoint[]`\n3. **Synthesizes** the profile and key points into a comprehensive `Summary` with abstract, ranked key points, themes, and open questions\n\n## Concepts\n\n- **DocumentProfile**: Metadata — `title`, `document_type`, `structure_type`, `language`, `sections_overview`\n- **KeyPoint**: Individual finding with `point`, `importance`, `category`, `supporting_evidence`\n- **Summary**: Final synthesis with `abstract`, `key_points[]`, `themes[]`, `open_questions[]`"
license = "MIT"
authors = ["Evotis S.A.S"]
main_pipe = "summarize_document"

[exports.document_summarization]
pipes = [
  "summarize_document",
  "profile_document",
  "extract_key_points",
  "synthesize_summary",
]