What can 한글 mcp hwpx MCP Server do?

stateless-hwpx-document-reading-with-context-preservation, hwpx-to-markdown-conversion-with-chunking, document-copying-with-safety-validation, table-creation-with-configurable-dimensions, table-cell-text-modification-with-range-support, table-cell-merging-and-splitting-for-complex-layouts, table-formatting-with-header-styling, text-range-formatting-with-multiple-attributes, custom-style-creation-and-management, document-style-listing-and-inspection, opc-package-introspection-with-part-enumeration, opc-part-xml-and-text-extraction, hwpx-to-html-conversion-with-semantic-preservation, xml-element-search-by-tag-and-attribute, edit-planning-with-preview-and-atomic-application, document-structure-validation-and-integrity-checking, text-convention-linting-for-style-consistency, mcp-server-transport-abstraction-with-stdio-and-http, hwpx-structure-extraction-to-json-with-toc-generation, keyword-search-with-context-window-retrieval, search-and-replace-with-split-run-preservation, paragraph-level-editing-with-insertion-and-deletion, heading-insertion-with-style-application, page-break-insertion-for-document-layout-control, memo-annotation-management-add-and-remove

한글 mcp hwpx MCP Server

MCP ServerFree

Open Source

/ 100

25 capabilities

Capabilities25 decomposed

stateless-hwpx-document-reading-with-context-preservation

Medium confidence

Reads HWPX (Open XML-based Korean word processor format) documents without requiring persistent state or the native Hangeul application, using pure Python XML parsing via python-hwpx library. Each tool invocation explicitly specifies the filename, enabling stateless operation across distributed MCP clients. The implementation parses OPC (Open Packaging Convention) package structure to extract document metadata, text content, and structural hierarchy (sections, paragraphs, tables) with character-level position tracking for context-aware retrieval.

Solves for

Extract full document text or metadata from HWPX files without installing Hangeul word processorRetrieve specific paragraph ranges with line/character positions for precise editing workflowsGet document outline and section structure to understand document hierarchy before processingList available HWPX documents in a folder to enable dynamic document selection in AI workflows

Best for

Korean enterprise automation teams migrating from Windows-only COM-based Hangeul automation

Cross-platform CI/CD pipelines requiring HWPX processing on Linux/macOS without GUI dependencies

AI agents (Claude, Gemini) that need to read Korean documents as context before editing

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path accessible to MCP server process

Limitations

Only supports HWPX (Open XML) format; legacy binary HWP files require conversion or separate tooling

Text extraction respects document structure but may lose complex formatting metadata (font families, exact spacing)

Character position tracking assumes UTF-8 encoding; mixed-encoding documents may have offset errors

What makes it unique

Implements stateless document reading where every tool call includes explicit filename parameter, eliminating session state management and enabling safe concurrent access from multiple MCP clients. Uses OPC package parsing to directly access document.xml without requiring Hangeul application or binary format reverse-engineering.

vs alternatives

Eliminates Windows/Hangeul dependency that COM-based automation requires, while maintaining full cross-platform support (Windows/macOS/Linux) with pure Python implementation.

hwpx-to-markdown-conversion-with-chunking

Medium confidence

Converts HWPX documents to Markdown format with configurable output strategies (full document or chunked by section/paragraph). The conversion pipeline parses document structure, maps heading levels to Markdown syntax (#, ##, ###), preserves table formatting as Markdown tables, and optionally splits output into chunks based on character limits or structural boundaries. Supports both file-based input (filename) and payload-based input (base64-encoded HWPX or HTTPS URL), enabling both local and remote document processing.

Solves for

Convert Korean HWPX documents to Markdown for version control, Git-friendly diffs, and documentation workflowsSplit large documents into Markdown chunks for LLM context windows (8000 chars default per chunk)Fetch HWPX from remote HTTPS URLs and convert on-the-fly without local file storageGenerate Markdown with metadata (section count, paragraph count, table count) for document analysis

Best for

Documentation teams converting legacy Hangeul documents to Markdown-based knowledge bases

LLM agents that need to chunk documents for RAG pipelines with section-aware boundaries

CI/CD workflows that require Markdown output for automated report generation

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL (exactly one required)

Limitations

Complex formatting (multi-column layouts, embedded objects, custom shapes) is simplified to Markdown equivalents

Images are referenced but not embedded in Markdown output; image extraction requires separate image-processing capability

Chunking strategy is structural (section/paragraph) only; semantic chunking based on content similarity not supported

What makes it unique

Supports dual input modes (file path + payload-based base64/URL) with identical tool interface, enabling both local document processing and remote document fetching. Chunking respects document structure (section/paragraph boundaries) rather than naive character splitting, preserving semantic coherence for LLM processing.

vs alternatives

Faster than Hangeul application export because it uses direct XML parsing; more flexible than pandoc because it understands HWPX-specific structure (sections, custom styles) natively.

document-copying-with-safety-validation

Medium confidence

Creates a safe copy of HWPX documents with validation to ensure copy integrity. The copy operation reads source document, validates structure, writes to destination path, and verifies copy completeness. Useful for creating backups before destructive operations or generating document variants.

Solves for

Create backup copies of documents before performing risky editsGenerate document variants by copying and modifyingEnsure document integrity by validating copy completeness

Best for

Document editing workflows that need safe backups before modifications

Batch document processing that generates variants from templates

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid source HWPX file path

Limitations

Copy validation is structural only; does not verify content accuracy

Large documents (>100MB) may have slow copy performance

Destination path must be writable; permission errors not handled gracefully

What makes it unique

Validates copy integrity by comparing source and destination structure, ensuring safe document duplication before risky operations.

vs alternatives

More reliable than OS-level file copy because it validates HWPX structure integrity; enables safe document variant generation.

table-creation-with-configurable-dimensions

Medium confidence

Creates new tables in HWPX documents with specified row and column counts. The table creation engine inserts a table element at specified paragraph index, initializes cells with empty content, and optionally applies basic formatting (header row styling). Tables are fully editable after creation via cell-level operations.

Solves for

Insert data tables into documents programmatically with specified dimensionsCreate table templates for data populationEnable AI agents to structure data in tabular format within documents

Best for

Document generation workflows that need to insert data tables

Report generation that presents data in tabular format

AI agents that structure data in tables

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Initial table cells are empty; content must be added via set_table_cell_text

Table dimensions are fixed at creation; resizing requires cell-level operations

Complex table structures (merged cells, nested tables) require post-creation manipulation

What makes it unique

Creates properly structured HWPX tables with configurable dimensions and optional header styling, enabling data-driven document generation.

vs alternatives

More flexible than template-based tables because dimensions are configurable; more reliable than manual table insertion because it maintains HWPX structure integrity.

table-cell-text-modification-with-range-support

Medium confidence

Modifies text content in table cells by table index, row, and column coordinates. The cell modification engine locates target cell, replaces content while preserving cell formatting, and supports batch operations for multiple cell updates. Enables data population in pre-created tables.

Solves for

Populate table cells with data programmaticallyUpdate specific cells in existing tablesEnable AI agents to fill tables with structured data

Best for

Data-driven document generation that populates tables with values

Report generation that fills template tables with computed data

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Cell content must be plain text; cannot insert formatted content

Cell coordinates (row, column) are 0-based; invalid coordinates fail silently

Cell text is replaced entirely; partial updates require read-modify-write

What makes it unique

Provides cell-level table editing with coordinate-based access, enabling precise data population in tables.

vs alternatives

More granular than document-level editing because it operates at cell level; enables efficient data-driven table population.

table-cell-merging-and-splitting-for-complex-layouts

Medium confidence

Manages table cell merging and splitting operations to create complex table layouts. The merge operation combines adjacent cells horizontally or vertically, and the split operation divides merged cells back into individual cells. Supports both single-cell and range-based operations.

Solves for

Create complex table layouts with merged cells for headers or grouped dataAdjust table structure by splitting merged cellsEnable AI agents to design table layouts programmatically

Best for

Report generation with complex table layouts (merged headers, grouped rows)

Template-based document generation with sophisticated table structures

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Merge operations are limited to adjacent cells; non-adjacent merges not supported

Split operations may lose content if merged cell contains text

Complex merge patterns may have unintended visual effects in rendered documents

What makes it unique

Supports both cell merging and splitting with directional control, enabling creation of complex table layouts programmatically.

vs alternatives

More flexible than fixed table structures because it allows dynamic layout adjustments; enables sophisticated report generation with merged headers and grouped data.

table-formatting-with-header-styling

Medium confidence

Applies basic formatting to tables including header row styling, border configuration, and cell alignment. The formatting engine applies predefined style templates to table elements, enabling consistent table appearance across documents.

Solves for

Apply professional formatting to tables for consistent document appearanceStyle table headers for visual distinctionEnable AI agents to create well-formatted tables

Best for

Document generation that requires consistent table formatting

Report generation with professional table styling

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Formatting options are limited to predefined templates; custom formatting requires direct XML manipulation

Formatting is applied at table level; cell-level custom formatting not supported

Color and font choices are limited to document style definitions

What makes it unique

Provides template-based table formatting that applies consistent styling across tables, enabling professional document appearance.

vs alternatives

More convenient than manual cell-by-cell formatting because it applies templates; ensures consistent table appearance across documents.

text-range-formatting-with-multiple-attributes

Medium confidence

Applies formatting attributes (bold, italic, underline, color, font size, font family) to text ranges within paragraphs. The formatting engine locates target text, applies formatting attributes to underlying XML runs, and preserves existing formatting for non-targeted attributes. Supports both single-range and multi-range formatting operations.

Solves for

Apply emphasis formatting (bold, italic, underline) to specific text within paragraphsChange text color or font size for visual distinctionEnable AI agents to apply formatting to generated or modified content

Best for

Document generation that needs formatted text (emphasis, color coding)

AI agents that apply formatting to highlight important content

Report generation with styled text elements

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Formatting is applied to exact text matches; partial word matches not supported

Font family and size changes may affect document layout; use with caution

Formatting attributes are limited to basic text properties; advanced formatting (shadows, effects) not supported

What makes it unique

Applies multiple formatting attributes simultaneously to text ranges while preserving existing formatting on non-targeted attributes.

vs alternatives

More flexible than paragraph-level formatting because it operates at text range level; enables selective emphasis within paragraphs.

custom-style-creation-and-management

Medium confidence

Creates and manages custom paragraph and character styles in HWPX documents. The style creation engine defines new styles with configurable properties (font, size, color, spacing), adds styles to document style sheet, and enables reuse across document. Supports style inheritance and override.

Solves for

Create custom styles for consistent formatting across documentsDefine organizational style standards (company heading style, body text style)Enable AI agents to apply custom styles to generated content

Best for

Enterprise document generation with organizational style standards

Template-based document creation that requires consistent styling

AI agents that apply custom styles to generated content

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Style creation requires detailed property specification; predefined templates not available

Style inheritance is limited to single-level inheritance; complex style hierarchies not supported

Style properties are limited to basic text formatting; advanced properties not supported

What makes it unique

Enables creation of custom styles that are stored in document style sheet and reusable across document, supporting organizational style standards.

vs alternatives

More flexible than predefined styles because custom styles can be tailored to organizational requirements; enables consistent branding across documents.

document-style-listing-and-inspection

Medium confidence

Lists all available styles in a document (paragraph and character styles) with detailed property inspection. The style listing engine enumerates styles from document style sheet, extracts style properties (font, size, color, spacing), and returns structured style information for inspection or reuse.

Solves for

Discover available styles in a document for reuseInspect style properties to understand document formatting standardsEnable AI agents to select appropriate styles for content

Best for

Document analysis workflows that need to understand style definitions

Style-aware document generation that reuses existing styles

Documentation of organizational style standards

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Style listing is read-only; does not modify styles

Style properties are extracted as-is; computed properties not calculated

Custom style properties may not be fully captured in listing

What makes it unique

Provides comprehensive style listing with property extraction, enabling style-aware document generation and analysis.

vs alternatives

More informative than simple style enumeration because it includes detailed property information; enables intelligent style selection for generated content.

opc-package-introspection-with-part-enumeration

Medium confidence

Inspects OPC (Open Packaging Convention) package structure of HWPX files, enumerating all package parts (document.xml, styles.xml, numbering.xml, etc.) with metadata. Available only in advanced mode (HWPX_MCP_ADVANCED=1). The introspection engine reads package manifest, lists parts with content types and relationships, enabling low-level document analysis.

Solves for

Analyze document package structure for debugging or advanced customizationDiscover available parts for direct XML manipulationEnable advanced users to understand document internals

Best for

Advanced users performing low-level document customization

Debugging document structure issues

Custom document processing that requires part-level access

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Part enumeration is read-only; does not modify package structure

Part relationships are complex; full relationship graph not always captured

What makes it unique

Provides low-level OPC package introspection for advanced users, enabling direct access to document internals for customization or debugging.

vs alternatives

More detailed than high-level document analysis because it exposes package structure; enables advanced customization not possible through standard tools.

opc-part-xml-and-text-extraction

Medium confidence

Extracts raw XML or text content from specific OPC package parts (document.xml, styles.xml, etc.). Available only in advanced mode (HWPX_MCP_ADVANCED=1). The extraction engine locates target part, reads content, and returns either raw XML or extracted text depending on part type.

Solves for

Extract raw XML from document parts for analysis or custom processingDebug document structure by inspecting underlying XMLEnable advanced users to perform custom XML transformations

Best for

Advanced users performing custom XML transformations

Debugging document structure issues

Custom document processing requiring direct XML access

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

XML extraction is read-only; modifications require separate apply_edit capability

Large XML parts may exceed response size limits

What makes it unique

Provides direct access to OPC part XML content for advanced users, enabling custom XML transformations and low-level document manipulation.

vs alternatives

More powerful than high-level APIs because it exposes raw XML; enables custom transformations not possible through standard tools.

hwpx-to-html-conversion-with-semantic-preservation

Medium confidence

Converts HWPX documents to HTML5 with semantic tag preservation (headings, sections, tables, lists). The conversion engine maps HWPX paragraph styles to HTML semantic elements, preserves table structure with proper thead/tbody/tfoot, and handles image references (with configurable image policy: omitted, base64-embedded, or URL-referenced). Supports chunked output with section-aware boundaries, enabling HTML fragments suitable for web embedding or LLM processing.

Solves for

Convert Korean documents to web-ready HTML for publishing to internal wikis or documentation sitesGenerate HTML chunks for LLM processing while preserving semantic structure (headings, sections)Extract document structure as semantic HTML for accessibility compliance and screen reader compatibilityEmbed HWPX content in web applications without requiring server-side Hangeul installation

Best for

Enterprise knowledge management systems converting legacy Hangeul documents to web-accessible HTML

Accessibility teams requiring semantic HTML output for WCAG compliance

LLM agents that need structured HTML for improved parsing and understanding of document hierarchy

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL

Limitations

Images are omitted by default (image_policy='omitted'); base64 embedding increases payload size significantly

CSS styling is minimal (semantic HTML only); complex formatting requires post-processing with CSS

Form fields and interactive elements in HWPX are converted to static HTML

What makes it unique

Generates semantic HTML5 with proper heading hierarchy and section structure derived from HWPX paragraph styles, rather than flat div-based output. Implements configurable image policy (omitted/base64/URL) to balance payload size vs. content completeness.

vs alternatives

Produces semantic HTML suitable for accessibility and SEO, unlike simple regex-based converters; preserves document structure better than generic XML-to-HTML transformations because it understands HWPX style semantics.

xml-element-search-by-tag-and-attribute

Medium confidence

Searches for XML elements in document parts by tag name or attribute values. Available only in advanced mode (HWPX_MCP_ADVANCED=1). The search engine queries document.xml or other parts using XPath-like patterns, returning matching elements with context.

Solves for

Find specific XML elements for custom processing or analysisLocate elements by attribute values (e.g., find all paragraphs with specific style)Enable advanced users to perform targeted XML queries

Best for

Advanced users performing custom XML queries

Document analysis that requires element-level search

Custom processing based on XML structure

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Search patterns are limited to simple tag/attribute matching; complex XPath not supported

Search results may be large for common tags; filtering required

What makes it unique

Provides element-level XML search by tag and attribute, enabling targeted queries on document structure.

vs alternatives

More precise than full-document XML parsing because it targets specific elements; enables efficient element-level analysis.

edit-planning-with-preview-and-atomic-application

Medium confidence

Provides a three-step edit workflow: plan_edit (describe desired changes), preview_edit (show what changes will be made), and apply_edit (execute changes atomically). Available only in advanced mode (HWPX_MCP_ADVANCED=1). The edit pipeline enables AI agents to propose changes, review them before application, and apply atomically with rollback capability.

Solves for

Plan complex document edits with AI assistance before applyingPreview changes to verify correctness before committingEnable safe, reversible document modifications with rollback capability

Best for

Complex document editing workflows requiring change review

AI-driven document modification with human oversight

High-risk document processing requiring change verification

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Preview is text-based; visual preview not available

Rollback requires manual restoration from backup; automatic rollback not supported

What makes it unique

Implements three-step edit workflow (plan → preview → apply) with atomic application, enabling safe document modification with change verification.

vs alternatives

More cautious than direct editing because it requires explicit preview and approval; enables AI-driven editing with human oversight.

document-structure-validation-and-integrity-checking

Medium confidence

Validates HWPX document structure for integrity and consistency. Available only in advanced mode (HWPX_MCP_ADVANCED=1). The validation engine checks for structural errors (malformed XML, missing required elements, broken references), reports issues with severity levels, and suggests corrections.

Solves for

Verify document integrity after modificationsDetect structural issues that may cause rendering problemsEnable document quality assurance workflows

Best for

Document quality assurance workflows

Post-modification validation to ensure document integrity

Debugging document structure issues

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Validation is structural only; does not check content accuracy

Validation may report false positives for non-standard but valid structures

What makes it unique

Provides comprehensive structure validation with severity levels and correction suggestions, enabling document quality assurance.

vs alternatives

More thorough than simple format checking because it validates structural consistency; enables early detection of issues that may cause rendering problems.

text-convention-linting-for-style-consistency

Medium confidence

Lints document text for style consistency violations (spacing, punctuation, naming conventions). Available only in advanced mode (HWPX_MCP_ADVANCED=1). The linting engine scans document paragraphs, checks against configurable style rules (e.g., spacing before punctuation, consistent terminology), and reports violations with suggestions.

Solves for

Enforce text style conventions across documentsDetect inconsistent terminology or formattingEnable document style quality assurance

Best for

Documentation teams enforcing style guidelines

Quality assurance workflows for consistent document style

Automated style checking for organizational standards

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Linting rules are predefined; custom rules not easily configurable

Korean language-specific rules may have false positives

What makes it unique

Provides text convention linting with configurable rules for style consistency, enabling automated quality assurance for document style.

vs alternatives

More specific than generic spell-checking because it focuses on style conventions; enables enforcement of organizational writing standards.

mcp-server-transport-abstraction-with-stdio-and-http

Medium confidence

Provides transport-agnostic MCP server implementation supporting both stdio (default) and streamable HTTP transports. The transport abstraction layer enables the same tool set to be accessed via stdio (for Claude Desktop, Cursor) or HTTP (for remote clients, web-based MCP clients). Transport selection via CLI arguments or environment variables (HWPX_MCP_TRANSPORT, HWPX_MCP_HOST, HWPX_MCP_PORT).

Solves for

Use hwpx-mcp-server with Claude Desktop or Cursor via stdio transportAccess hwpx-mcp-server from remote clients or web applications via HTTPDeploy hwpx-mcp-server in containerized environments with HTTP transport

Best for

Teams using Claude Desktop or Cursor that prefer stdio transport

Remote document processing workflows requiring HTTP access

Containerized deployments where HTTP is preferred over stdio

Requires

Python >= 3.10

python-hwpx >= 1.9

For HTTP: accessible network port (default 8000)

Limitations

HTTP transport lacks authentication (development-focused); not suitable for untrusted networks

Stdio transport limited to single client connection; HTTP supports multiple concurrent clients

HTTP transport adds ~50-100ms latency per request compared to stdio

What makes it unique

Implements transport abstraction that enables identical tool set to be accessed via stdio or HTTP, supporting diverse deployment scenarios without code duplication.

vs alternatives

More flexible than single-transport servers because it supports both stdio and HTTP; enables use in both desktop (Claude) and server (remote) environments.

hwpx-structure-extraction-to-json-with-toc-generation

Medium confidence

Extracts complete HWPX document structure (title, table of contents, sections, paragraphs, tables, figures) into structured JSON format with automatic TOC generation. The extraction engine parses document.xml to identify heading styles, builds hierarchical section tree with paragraph indices, extracts table metadata (row/column counts, cell content), and catalogs figure captions. Supports chunked output with metadata for each chunk (chunk_index, strategy, section reference).

Solves for

Extract document structure for programmatic analysis (section count, nesting depth, table inventory)Generate table of contents automatically from heading styles for navigation or outline generationBuild document index for semantic search or RAG systems that need paragraph-level granularityAnalyze document complexity (table count, figure count, nesting depth) for workflow routing

Best for

Knowledge management systems building searchable indices of Korean documents

RAG pipelines that need structured document metadata for retrieval ranking

Document analysis tools that classify documents by complexity or structure type

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL

Limitations

TOC generation relies on heading style detection; documents with inconsistent heading styles may produce incomplete TOC

Figure extraction captures captions only; image content itself is not extracted (use image-processing capability separately)

Nested table structures are flattened to single-level table list; complex table hierarchies lose nesting information

What makes it unique

Automatically generates hierarchical table of contents from heading styles with paragraph index references, enabling direct navigation to specific paragraphs. Produces structured JSON with explicit section nesting and chunk metadata, suitable for RAG indexing and semantic search.

vs alternatives

More structured than simple text extraction because it preserves document hierarchy and metadata; more accurate than regex-based heading detection because it uses HWPX style information directly.

keyword-search-with-context-window-retrieval

Medium confidence

Searches HWPX documents for keyword matches and returns results with configurable context windows (surrounding paragraphs). The search engine iterates through document paragraphs, performs case-insensitive substring matching, and returns match location (paragraph index, character offset) with surrounding context (previous/next N paragraphs). Supports regex patterns and exact phrase matching via parameter configuration.

Solves for

Find all occurrences of a keyword in a document and retrieve surrounding context for AI analysisLocate specific sections by searching for section-identifying keywordsBuild search indices for document discovery without full-text search infrastructureEnable AI agents to navigate documents by searching for relevant content before editing

Best for

Document review workflows where AI needs to locate and analyze specific content

Search-based navigation in large Korean documents without external search engines

Content discovery for RAG systems that need paragraph-level search results

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Search is linear (O(n) complexity); large documents (>10000 paragraphs) may have noticeable latency

Context window is paragraph-based only; character-level context not supported

Regex patterns require careful escaping; complex regex may have performance impact

What makes it unique

Returns match location with paragraph-level context windows, enabling AI agents to understand surrounding content without loading entire document. Supports both substring and regex matching with configurable context depth.

vs alternatives

Faster than full-document loading for targeted searches; more context-aware than simple find-and-replace because it returns surrounding paragraphs for semantic understanding.

search-and-replace-with-split-run-preservation

Medium confidence

Performs single or batch text replacement in HWPX documents with preservation of split-run formatting (text runs that span multiple formatting regions). The replacement engine locates target text, handles cases where text spans multiple XML runs (common in formatted documents), reconstructs runs to maintain formatting integrity, and saves changes back to the HWPX file. Batch mode supports multiple replacements in a single operation with atomic save.

Solves for

Replace specific text in Korean documents while preserving bold/italic/color formattingPerform bulk find-and-replace operations (e.g., update all references to a name or term)Automate document template population by replacing placeholder text with actual valuesEnable AI agents to modify document content without corrupting formatting metadata

Best for

Document template automation where placeholders need replacement with formatted content

Bulk document updates (e.g., company name change across multiple documents)

AI-driven document editing workflows that need to preserve formatting

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Split-run handling adds complexity; replacements spanning multiple runs may fail silently if run structure is malformed

Replacement text must be plain text; cannot insert formatted text (use add_paragraph for formatted content)

Case-sensitive matching only; case-insensitive replacement requires pre-processing

What makes it unique

Implements split-run aware replacement that reconstructs XML runs to preserve formatting when replacement text spans multiple formatting regions. Supports batch replacements with atomic save, enabling multi-step document transformations.

vs alternatives

Preserves formatting metadata that naive string replacement would corrupt; more reliable than regex-based replacement because it understands HWPX XML structure.

paragraph-level-editing-with-insertion-and-deletion

Medium confidence

Provides granular paragraph-level editing operations: add new paragraphs (with optional style), insert paragraphs at specific indices, and delete paragraphs by index or range. The editing engine manipulates document.xml to insert/remove paragraph elements while maintaining document structure integrity, updates paragraph indices in cross-references, and optionally applies paragraph styles (heading, normal, list, etc.). Each operation is atomic and includes optional backup creation.

Solves for

Add new content paragraphs to documents programmatically with specific styles (heading, body, list)Insert paragraphs at specific locations for structured document buildingDelete obsolete paragraphs or sections by indexEnable AI agents to build documents incrementally by adding paragraphs one at a time

Best for

Document generation workflows that build documents from templates by adding paragraphs

AI agents that need to modify document structure (add sections, delete outdated content)

Automated report generation that assembles documents from content blocks

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Paragraph indices are 0-based and may shift after insertions/deletions; subsequent operations must use updated indices

Style application is limited to predefined styles; custom style creation requires separate capability

Paragraph content must be plain text; formatted text requires separate formatting capability

What makes it unique

Provides index-based paragraph manipulation with style application, enabling precise document structure editing. Maintains document integrity by updating internal cross-references after insertions/deletions.

vs alternatives

More precise than document-level editing because it operates at paragraph granularity; more flexible than template-based approaches because it supports arbitrary insertion/deletion at any index.

heading-insertion-with-style-application

Medium confidence

Specialized paragraph insertion for adding heading elements with automatic style application (Heading 1, Heading 2, etc.). The operation inserts a new paragraph with specified heading level, applies corresponding style from document style sheet, and optionally adjusts surrounding paragraph styles for proper hierarchy. Supports insertion at specific indices or at document end.

Solves for

Add section headings to documents programmatically with proper style hierarchyCreate document structure by inserting headings at appropriate nesting levelsEnable AI agents to organize document content by adding section headersBuild table of contents by inserting heading hierarchy

Best for

Document generation workflows that need to create structured documents with proper heading hierarchy

AI agents that organize content by adding section headers

Automated report generation that creates section structure

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Heading style must exist in document style sheet; custom heading styles require separate style creation

Heading level (1-6) must be valid; invalid levels may default to Heading 1

Heading hierarchy validation is not enforced; can create Heading 3 without Heading 1/2

What makes it unique

Specialized heading insertion that automatically applies correct heading style based on level parameter, ensuring proper document hierarchy and TOC generation.

vs alternatives

More convenient than generic paragraph insertion because it handles style application automatically; ensures proper heading hierarchy for TOC generation.

page-break-insertion-for-document-layout-control

Medium confidence

Inserts page break elements into HWPX documents at specified paragraph indices, enabling control over document pagination and layout. The operation inserts a page break marker (special paragraph element) that forces subsequent content to a new page, useful for section separation or template-based document generation.

Solves for

Insert page breaks between document sections for proper layoutControl pagination in generated documents (e.g., each report section on new page)Enable AI agents to structure multi-section documents with proper page separation

Best for

Document generation workflows that need precise pagination control

Template-based document building where sections must start on new pages

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Page break insertion is position-based; actual page layout depends on content size and margins

Multiple consecutive page breaks may create blank pages

Page break behavior may vary depending on document margins and font sizes

What makes it unique

Provides explicit page break insertion at paragraph level, enabling precise pagination control in programmatically generated documents.

vs alternatives

More reliable than relying on content size for pagination; enables predictable document layout in template-based generation.

memo-annotation-management-add-and-remove

Medium confidence

Manages memo (comment/annotation) elements in HWPX documents with add and remove operations. The memo engine inserts memo markers at specified paragraph indices with configurable memo text, and removes existing memos by index. Memos are preserved in document metadata and visible in Hangeul application when document is opened.

Solves for

Add review comments or annotations to documents programmaticallyEnable AI agents to mark content for human review with contextual notesRemove outdated annotations or memos from documents

Best for

Document review workflows where AI adds comments for human review

Collaborative editing where AI annotations guide human editors

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Memo text is plain text only; cannot include formatted content

Memo visibility depends on Hangeul application version; may not display in all viewers

Memo indices may shift after paragraph insertions/deletions

What makes it unique

Provides programmatic memo management that integrates with Hangeul application's native annotation system, enabling AI-driven document review workflows.

vs alternatives

More integrated than external annotation systems because memos are stored in document metadata and visible in Hangeul; enables collaborative workflows between AI and human reviewers.

Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.

Related Artifactssharing capabilities

Artifacts that share capabilities with 한글 mcp hwpx MCP Server , ranked by overlap. Discovered automatically through the match graph.

Framework22

privateGPT

Ask questions to your documents without an internet connection, using the power of LLMs.

document-chunking-and-context-windowing

1 shared capability

Framework59

Haystack

Production NLP/LLM framework for search and RAG pipelines with component-based architecture.

document processing pipeline with format conversion and chunking

1 shared capability

CLI Tool21

llama-parse

Parse files into RAG-Optimized formats.

semantic document chunking with context preservation

1 shared capability

Product51

Hebbia

Revolutionize document analysis: AI collaboration, transparency, vast data...

complex document format preservation

1 shared capability

Repository22

Local GPT

Chat with documents without compromising privacy

multi-format-document-ingestion-with-contextual-enrichment

1 shared capability

MCP Server38

rag-memory-epf-mcp

Project-local RAG memory MCP server — knowledge graph + multilingual vector + FTS5 in a single SQLite file. Per-project isolation, 30 MCP tools, codepoint-safe chunking (Korean/CJK/emoji).

semantic chunking with context preservation

1 shared capability

Best For

✓Korean enterprise automation teams migrating from Windows-only COM-based Hangeul automation
✓Cross-platform CI/CD pipelines requiring HWPX processing on Linux/macOS without GUI dependencies
✓AI agents (Claude, Gemini) that need to read Korean documents as context before editing
✓Documentation teams converting legacy Hangeul documents to Markdown-based knowledge bases
✓LLM agents that need to chunk documents for RAG pipelines with section-aware boundaries
✓CI/CD workflows that require Markdown output for automated report generation
✓Document editing workflows that need safe backups before modifications
✓Batch document processing that generates variants from templates

Known Limitations

⚠Only supports HWPX (Open XML) format; legacy binary HWP files require conversion or separate tooling
⚠Text extraction respects document structure but may lose complex formatting metadata (font families, exact spacing)
⚠Character position tracking assumes UTF-8 encoding; mixed-encoding documents may have offset errors
⚠Maximum character return limited by HWPX_MCP_MAX_CHARS env var (default 10000) to prevent context overflow
⚠Complex formatting (multi-column layouts, embedded objects, custom shapes) is simplified to Markdown equivalents
⚠Images are referenced but not embedded in Markdown output; image extraction requires separate image-processing capability

Requirements

Python >= 3.10python-hwpx >= 1.9Valid HWPX file path accessible to MCP server processEither valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL (exactly one required)Valid source HWPX file pathWrite permissions for destination pathValid HWPX file path with write permissionsTable must exist at specified index

Input / Output

Accepts: file_path (string, relative or absolute), paragraph_index (integer), paragraph_range (start_index, end_index integers), filename (string, file path), hwpx_base64 (base64-encoded binary HWPX content), url (HTTPS URL string pointing to HWPX file), output (enum: 'full' or 'chunks'), chunk_strategy (enum: 'section' or 'paragraph'), max_chars_per_chunk (integer, default from HWPX_MCP_MAX_CHARS_PER_CHUNK env var), filename (string, source file path), destination_path (string, target file path), rows (integer, number of rows), columns (integer, number of columns), index (integer, optional insertion position), header_row (boolean, optional, apply header styling to first row), table_index (integer, table identifier), row (integer, 0-based row index), column (integer, 0-based column index), text (string, cell content), start_row (integer, starting row for merge), start_column (integer, starting column for merge), end_row (integer, ending row for merge), end_column (integer, ending column for merge), direction (enum: 'horizontal' or 'vertical'), style_template (enum: 'default', 'professional', 'minimal'), header_style (enum: 'bold', 'shaded', 'bordered'), paragraph_index (integer, target paragraph), text (string, text to format), bold (boolean, optional), italic (boolean, optional), underline (boolean, optional), color (string, optional, hex color code), font_size (integer, optional, in points), font_family (string, optional), style_name (string, custom style identifier), style_type (enum: 'paragraph' or 'character'), base_style (string, optional, parent style for inheritance), properties (object with font, size, color, spacing, etc.), style_type (enum: 'all', 'paragraph', 'character', optional), part_name (string, OPC part identifier, e.g., 'word/document.xml'), output_format (enum: 'xml' or 'text'), url (HTTPS URL string), image_policy (enum: 'omitted', 'base64', 'url-referenced'), max_chars_per_chunk (integer), tag_name (string, XML tag to search for), attribute_name (string, optional, attribute to filter by), attribute_value (string, optional, attribute value to match), edit_description (string, description of desired changes), edit_plan (object, structured edit plan from plan_edit), apply (boolean, whether to apply changes), strict_mode (boolean, optional, enable strict validation), rules (array of rule identifiers, optional), transport (enum: 'stdio' or 'streamable-http', via CLI or env var), host (string, HTTP bind address, default 127.0.0.1), port (integer, HTTP bind port, default 8000), keyword (string, search term), context_paragraphs (integer, number of surrounding paragraphs to include), case_sensitive (boolean, default false), regex_mode (boolean, default false), find_text (string, text to search for), replace_text (string, replacement text), replacements (array of {find_text, replace_text} objects for batch mode), text (string, paragraph content), style (string, paragraph style name, e.g., 'Heading 1', 'Normal'), index (integer, insertion position for insert_paragraph), paragraph_index (integer, target paragraph for deletion), text (string, heading text), level (integer, 1-6 for Heading 1-6), index (integer, paragraph index where page break should be inserted), paragraph_index (integer, target paragraph for memo), memo_text (string, memo content), memo_index (integer, memo identifier for removal)

Produces: plain_text (UTF-8 string), structured_json (document metadata with section/paragraph/table counts), outline_json (hierarchical heading structure with paragraph indices), markdown (string, full Markdown document or single chunk), chunks (array of Markdown strings), meta (object with source_type, section_count, paragraph_count, table_count, figure_caption_count), destination_path (string, path to copied file), status (string, 'success' or error message), validation_result (object with structure_valid, size_match), table_index (integer, index of created table), cell_reference (string, e.g., 'Table 0, Row 1, Col 2'), merged_cells (integer, number of cells merged), formatting_applied (object with header_styled, borders_applied), formatting_applied (object with attributes applied), style_id (string, internal style identifier), styles (array of style objects with name, type, properties), style_count (integer, total styles in document), parts (array of part objects with name, content_type, relationships), part_count (integer, total parts in package), content (string, raw XML or extracted text), part_name (string, confirmed part identifier), content_type (string, MIME type), html (string, full HTML document or single chunk), chunks (array of HTML fragment strings), meta (object with source_type, image_policy, chunk_count), elements (array of matching XML elements with context), match_count (integer, total matches found), plan (object, structured edit plan with operations), preview (string, text preview of changes), changes_applied (integer, number of changes applied), valid (boolean, document is valid), issues (array of validation issues with severity and description), issue_count (integer, total issues found), violations (array of style violations with location and suggestion), violation_count (integer, total violations found), server_status (string, 'listening' or error message), transport_info (object with transport type, host, port), doc (object with title, toc array, sections array, tables array, figures array), chunks (array of chunk objects with chunk_index, strategy, section reference), meta (object with source_type, extraction_timestamp), matches (array of match objects with paragraph_index, character_offset, matched_text, context_before, context_after), replaced_count (integer, number of replacements made), backup_path (string, path to .bak file if HWPX_MCP_AUTOBACKUP=1), new_paragraph_index (integer, index of newly added paragraph), backup_path (string, path to .bak file if backup created), new_heading_index (integer, index of newly added heading), page_break_index (integer, index of inserted page break), memo_id (string, identifier for added memo)

UnfragileRank

Adoption5%(25% weight)

Quality50%(25% weight)

Ecosystem69%(15% weight)

Match Graph25%(30% weight)

Freshness75%(5% weight)

UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.

Type: MCP Server

25 capabilities

Visit 한글 mcp hwpx MCP Server →

Repository Details

About

<h1 align="center">📄 hwpx-mcp-server</h1> 한글(HWPX) 문서를 AI로 자동화하는 MCP 서버 한글 워드프로세서 없이 · 순수 파이썬 · 크로스 플랫폼 <a href="https://pypi.org/project/hwpx-mcp-server/"><img src="https://img.shields.io/pypi/v/hwpx-mcp-server?style=flat-square&color=blue" alt="PyPI"></a> <a href="https://pypi.org/project/hwpx-mcp-server/"><img src="https://img.shields.io/pypi/pyversions/hwpx-mcp-server?style=flat-square" alt="Python"></a> <a href="https://github.com/airmang/hwpx-mcp-server/blob/main/LICENSE"><img src="https://img.shields.io/github/license/airmang/hwpx-mcp-server?style=flat-square" alt="License"></a> <a href="https://github.com/airmang/hwpx-mcp-server/actions"><img src="https://img.shields.io/github/actions/workflow/status/airmang/hwpx-mcp-server/test.yml?style=flat-square&label=tests" alt="Tests"></a> --- **hwpx-mcp-server**는 [Model Context Protocol(MCP)](https://modelcontextprotocol.io) 표준을 따르는 서버로, [python-hwpx](https://github.com/airmang/python-hwpx) 기반에서 HWPX 문서의 열람 · 검색 · 편집 · 저장을 AI 클라이언트에서 직접 수행할 수 있게 합니다. > **Note** — 이 서버는 Open XML 기반 `.hwpx` 포맷을 지원합니다. 레거시 바이너리 `.hwp` 포맷은 직접 편집 대상이 아닙니다. ## Why? 국내 공공기관·학교·기업에서는 한글 문서 기반 업무가 매우 많지만, 자동화는 오랫동안 OS/프로그램 의존성이 컸습니다. **hwpx-mcp-server**는 이 제약을 줄이는 데 초점을 맞춥니다. - ✅ **OS 무관** — Windows, macOS, Linux에서 동작 - ✅ **한글 워드프로세서 불필요** — 순수 파이썬 기반 처리 - ✅ **AI 네이티브** — Claude Desktop, VS Code, Gemini CLI 등 MCP 클라이언트와 직접 연결 - ✅ **Stateless 기본 설계** — 도구 호출마다 `filename`을 명시해 일관적으로 실행 ## Use Cases - 실전 사용 사례 9개 보기: [`docs/use-cases.md`](docs/use-cases.md) - 종합 테스트 리포트: [`tests/hwpx_mcp_report_updated.md`](tests/hwpx_mcp_report_updated.md) ## Quick Start ### 1. 설치 & 실행 [uv](https://docs.astral.sh/uv/getting-started/installation/) 기준: ```bash uvx hwpx-mcp-server ``` 또는 pip 설치: ```bash pip install hwpx-mcp-server hwpx-mcp-server ``` 요구사항: - `Python >= 3.10` - `python-hwpx >= 1.9` ### 2. MCP 클라이언트 설정 <details> <summary>Claude Desktop</summary> `claude_desktop_config.json`: ```json { "mcpServers": { "hwpx": { "command": "uvx", "args": ["hwpx-mcp-server"] } } } ``` </details> <details> <summary>Gemini CLI</summary> `~/.gemini/settings.json`: ```json { "mcpServers": { "hwpx": { "command": "uvx", "args": ["hwpx-mcp-server"] } } } ``` </details> <details> <summary>VS Code (Copilot Chat)</summary> `.vscode/mcp.json`: ```json { "servers": { "hwpx": { "command": "uvx", "args": ["hwpx-mcp-server"] } } } ``` </details> <details> <summary>Cursor / Windsurf</summary> 각 에디터 MCP 설정 파일에 동일한 블록을 추가: ```json { "mcpServers": { "hwpx": { "command": "uvx", "args": ["hwpx-mcp-server"] } } } ``` </details> ### 3. 전송 모드 선택 (Stdio + Streamable HTTP) 기본 `stdio` 사용은 기존과 동일합니다. ```bash hwpx-mcp-server ``` 동일한 MCP 도구 세트를 Streamable HTTP로 실행할 수 있습니다. ```bash hwpx-mcp-server --transport streamable-http --host 127.0.0.1 --port 8000 ``` 환경 변수로도 동일하게 제어할 수 있습니다. - `HWPX_MCP_TRANSPORT` (`stdio` 또는 `streamable-http`) - `HWPX_MCP_HOST` (기본값: `127.0.0.1`) - `HWPX_MCP_PORT` (기본값: `8000`) > 참고: HTTP 인증은 현재 개발 편의 중심으로 단순하게 유지되어 있습니다. 프로덕션용 인증 훅은 서버 진입점에 TODO로 남겨두었습니다. ## Features 기본 모드에서 30개 도구, 고급 모드(`HWPX_MCP_ADVANCED=1`)에서 추가 10개 도구가 활성화됩니다. ### 📖 읽기 & 탐색 | 도구 | 설명 | |---|---| | `get_document_info` | 문서 메타데이터/섹션/문단/표 개수 조회 | | `get_document_text` | 문서 전체 텍스트 추출 (`max_chars` 지원) | | `get_document_outline` | 제목/개요 구조 추출 | | `get_paragraph_text` | 특정 문단 텍스트 조회 | | `get_paragraphs_text` | 문단 범위 조회 | | `list_available_documents` | 폴더 내 `.hwpx` 파일 목록 조회 | ### 🧾 변환 & 추출 (입력 페이로드 기반) | 도구 | 설명 | |---|---| | `hwpx_to_markdown` | HWPX 입력을 Markdown으로 변환 | | `hwpx_to_html` | HWPX 입력을 HTML로 변환 | | `hwpx_extract_json` | HWPX 구조를 JSON으로 추출 | 공통 입력 규칙: - 입력 소스는 `hwpx_base64` 또는 `url` 중 정확히 하나만 허용 - `url`은 `https://...`만 허용 공통 옵션: - `output`: `full` 또는 `chunks` - `chunk_strategy`: `section` 또는 `paragraph` - `max_chars_per_chunk`: 청크당 최대 문자 수(기본값: `HWPX_MCP_MAX_CHARS_PER_CHUNK` 또는 `8000`) <details> <summary>변환/추출 응답 예시</summary> #### `hwpx_to_markdown` ```json { "markdown": "# Title\n\nParagraph...", "chunks": ["..."], "meta": { "source_type": "base64", "section_count": 2, "paragraph_count": 10, "table_count": 1, "figure_caption_count": 1 } } ``` #### `hwpx_to_html` ```json { "html": "<!doctype html><html>...</html>", "chunks": ["<section>...</section>"], "meta": { "source_type": "url", "image_policy": "omitted" } } ``` #### `hwpx_extract_json` ```json { "doc": { "title": "Title", "toc": [{ "level": 1, "text": "Title", "paragraph_index": 0 }], "sections": [{ "index": 0, "title": "Title", "paragraphs": [] }], "tables": [], "figures": [] }, "chunks": [{ "chunk_index": 0, "strategy": "section", "section": {} }], "meta": { "source_type": "base64" } } ``` </details> ### 🔎 검색 & 치환 | 도구 | 설명 | |---|---| | `find_text` | 키워드 검색 + 컨텍스트 반환 | | `search_and_replace` | 단일 치환 (split-run 보강) | | `batch_replace` | 다중 치환 일괄 실행 | ### ✏️ 편집 | 도구 | 설명 | |---|---| | `add_heading` | 제목(헤딩) 문단 추가 | | `add_paragraph` / `insert_paragraph` / `delete_paragraph` | 문단 추가/삽입/삭제 | | `add_page_break` | 페이지 나누기 추가 | | `add_memo` / `remove_memo` | 메모 추가/제거 | | `copy_document` | 문서 안전 복사 | ### 📊 표 | 도구 | 설명 | |---|---| | `add_table` / `get_table_text` | 표 생성/조회 | | `set_table_cell_text` | 셀 텍스트 수정 | | `merge_table_cells` / `split_table_cell` | 셀 병합/분할 | | `format_table` | 표 헤더 등 기본 서식 적용 | ### 🎨 스타일 | 도구 | 설명 | |---|---| | `format_text` | 텍스트 범위 서식 적용(굵기, 기울임, 밑줄, 색상 등) | | `create_custom_style` | 커스텀 스타일 생성 | | `list_styles` | 문서 스타일 목록 조회 | ### 🔬 고급 (옵션) `HWPX_MCP_ADVANCED=1`일 때 활성화: | 도구 | 설명 | |---|---| | `package_parts` | OPC 파트 목록 조회 | | `package_get_xml` / `package_get_text` | 파트 XML/텍스트 조회 | | `object_find_by_tag` / `object_find_by_attr` | XML 요소 검색 | | `plan_edit` / `preview_edit` / `apply_edit` | 편집 계획/미리보기/적용 | | `validate_structure` / `lint_text_conventions` | 구조 검증/텍스트 린트 | ## Configuration | 변수 | 설명 | 기본값 | |---|---|---| | `HWPX_MCP_MAX_CHARS` | 텍스트 반환 도구 기본 최대 길이 | `10000` | | `HWPX_MCP_MAX_CHARS_PER_CHUNK` | 변환/추출 도구의 청크 분할 기본 길이 | `8000` | | `HWPX_MCP_AUTOBACKUP` | `1`이면 저장 전 `.bak` 백업 생성 | `1` | | `HWPX_MCP_ADVANCED` | `1`이면 고급 도구 활성화 | `0` | | `HWPX_MCP_TRANSPORT` | 서버 전송 모드 (`stdio`, `streamable-http`) | `stdio` | | `HWPX_MCP_HOST` | HTTP 바인딩 호스트 | `127.0.0.1` | | `HWPX_MCP_PORT` | HTTP 바인딩 포트 | `8000` | | `LOG_LEVEL` | 로그 레벨 | `INFO` | 환경 변수 포함 MCP 설정 예시: ```json { "mcpServers": { "hwpx": { "command": "uvx", "args": ["hwpx-mcp-server"], "env": { "HWPX_MCP_MAX_CHARS": "12000", "HWPX_MCP_MAX_CHARS_PER_CHUNK": "8000", "HWPX_MCP_AUTOBACKUP": "1", "HWPX_MCP_ADVANCED": "0", "HWPX_MCP_TRANSPORT": "stdio", "HWPX_MCP_HOST": "127.0.0.1", "HWPX_MCP_PORT": "8000", "LOG_LEVEL": "INFO" } } } } ``` ## Advanced <details> <summary>📦 OPC 파트 조회</summary> 고급 모드에서 문서 내부 파트를 직접 조회할 수 있습니다. - `package_parts` - `package_get_xml` - `package_get_text` </details> <details> <summary>🧭 편집 파이프라인</summary> 고급 모드에서 `plan_edit → preview_edit → apply_edit` 흐름으로 변경 계획을 검토하고 적용할 수 있습니다. </details> <details> <summary>🧪 구조/규칙 검사</summary> 고급 모드에서 다음 검사 도구를 사용할 수 있습니다. - `validate_structure` - `lint_text_conventions` </details> ## Testing ```bash # 테스트 의존성 설치 python -m pip install -e ".[test]" # 전체 테스트 python -m pytest -q ``` 로컬 기준(2026-02-22) 전체 테스트가 통과했습니다. - 실전 사용 사례: `docs/use-cases.md` - 종합 리포트: `tests/hwpx_mcp_report_updated.md` - 회귀 테스트: `tests/test_hwpx_report_regressions.py` ## Architecture ```text hwpx-mcp-server ├── src/hwpx_mcp_server/ │ ├── server.py # Stateless MCP 진입점 │ ├── hwpx_ops.py # 고급/내부 연산 래퍼 │ ├── core/ # 문단/표/검색/서식 핵심 로직 │ ├── tools.py # 확장 도구 스키마/정의 │ └── schema/ # JSON 스키마 빌더/정리기 ├── tests/ # 단위 + E2E + 회귀 테스트 └── pyproject.toml ``` ## Comparison | | hwpx-mcp-server | hwp(바이너리) COM 자동화 계열 | |---|---|---| | 대상 포맷 | `.hwpx` (Open XML) | `.hwp` (바이너리) 중심 | | OS | Windows · macOS · Linux | 대체로 Windows 중심 | | 한글 프로그램 의존 | 불필요 | 필요한 경우가 많음 | | 연동 방식 | MCP + 파이썬 라이브러리 | 데스크톱 앱 자동화 | ## Contributing 기여를 환영합니다. 1. Fork 후 브랜치 생성 2. 변경 + 테스트 추가/수정 3. `pytest -q` 통과 확인 후 PR ## License [MIT](LICENSE) © 고규현 (Kyuhyun Koh) ## Author **고규현** — 광교고등학교 정보·컴퓨터 교사 - ✉️ [kokyuhyun@hotmail.com](mailto:kokyuhyun@hotmail.com) - 🐙 [@airmang](https://github.com/airmang)

Alternatives to 한글 mcp hwpx MCP Server

Supabase69Platform

Search the Supabase docs for up-to-date guidance and troubleshoot errors quickly. Manage organizations, projects, databases, and Edge Functions, including migrations, SQL, logs, advisors, keys, and type generation, in one flow. Create and manage development branches to iterate safely, confirm costs

Compare →

Tavily MCP Server62MCP Server

AI-optimized web search and content extraction via Tavily MCP.

Compare →

MongoDB MCP Server62MCP Server

Query and manage MongoDB databases and collections via MCP.

Compare →

Firecrawl MCP Server62MCP Server

Scrape websites and extract structured data via Firecrawl MCP.

Compare →

Are you the builder of 한글 mcp hwpx MCP Server ?

Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.

Claim this artifact →Verification via email

Get the weekly brief

New tools, rising stars, and what's actually worth your time. No spam.

Data Sources

smithery

Looking for something else?

Search →

Capabilities25 decomposed

stateless-hwpx-document-reading-with-context-preservation

Medium confidence

Solves for

Best for

Korean enterprise automation teams migrating from Windows-only COM-based Hangeul automation

Cross-platform CI/CD pipelines requiring HWPX processing on Linux/macOS without GUI dependencies

AI agents (Claude, Gemini) that need to read Korean documents as context before editing

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path accessible to MCP server process

Limitations

Only supports HWPX (Open XML) format; legacy binary HWP files require conversion or separate tooling

Text extraction respects document structure but may lose complex formatting metadata (font families, exact spacing)

Character position tracking assumes UTF-8 encoding; mixed-encoding documents may have offset errors

What makes it unique

vs alternatives

Eliminates Windows/Hangeul dependency that COM-based automation requires, while maintaining full cross-platform support (Windows/macOS/Linux) with pure Python implementation.

hwpx-to-markdown-conversion-with-chunking

Medium confidence

Solves for

Best for

Documentation teams converting legacy Hangeul documents to Markdown-based knowledge bases

LLM agents that need to chunk documents for RAG pipelines with section-aware boundaries

CI/CD workflows that require Markdown output for automated report generation

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL (exactly one required)

Limitations

Complex formatting (multi-column layouts, embedded objects, custom shapes) is simplified to Markdown equivalents

Images are referenced but not embedded in Markdown output; image extraction requires separate image-processing capability

Chunking strategy is structural (section/paragraph) only; semantic chunking based on content similarity not supported

What makes it unique

vs alternatives

Faster than Hangeul application export because it uses direct XML parsing; more flexible than pandoc because it understands HWPX-specific structure (sections, custom styles) natively.

document-copying-with-safety-validation

Medium confidence

Solves for

Create backup copies of documents before performing risky editsGenerate document variants by copying and modifyingEnsure document integrity by validating copy completeness

Best for

Document editing workflows that need safe backups before modifications

Batch document processing that generates variants from templates

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid source HWPX file path

Limitations

Copy validation is structural only; does not verify content accuracy

Large documents (>100MB) may have slow copy performance

Destination path must be writable; permission errors not handled gracefully

What makes it unique

Validates copy integrity by comparing source and destination structure, ensuring safe document duplication before risky operations.

vs alternatives

More reliable than OS-level file copy because it validates HWPX structure integrity; enables safe document variant generation.

table-creation-with-configurable-dimensions

Medium confidence

Solves for

Insert data tables into documents programmatically with specified dimensionsCreate table templates for data populationEnable AI agents to structure data in tabular format within documents

Best for

Document generation workflows that need to insert data tables

Report generation that presents data in tabular format

AI agents that structure data in tables

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Initial table cells are empty; content must be added via set_table_cell_text

Table dimensions are fixed at creation; resizing requires cell-level operations

Complex table structures (merged cells, nested tables) require post-creation manipulation

What makes it unique

Creates properly structured HWPX tables with configurable dimensions and optional header styling, enabling data-driven document generation.

vs alternatives

More flexible than template-based tables because dimensions are configurable; more reliable than manual table insertion because it maintains HWPX structure integrity.

table-cell-text-modification-with-range-support

Medium confidence

Solves for

Populate table cells with data programmaticallyUpdate specific cells in existing tablesEnable AI agents to fill tables with structured data

Best for

Data-driven document generation that populates tables with values

Report generation that fills template tables with computed data

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Cell content must be plain text; cannot insert formatted content

Cell coordinates (row, column) are 0-based; invalid coordinates fail silently

Cell text is replaced entirely; partial updates require read-modify-write

What makes it unique

Provides cell-level table editing with coordinate-based access, enabling precise data population in tables.

vs alternatives

More granular than document-level editing because it operates at cell level; enables efficient data-driven table population.

table-cell-merging-and-splitting-for-complex-layouts

Medium confidence

Solves for

Create complex table layouts with merged cells for headers or grouped dataAdjust table structure by splitting merged cellsEnable AI agents to design table layouts programmatically

Best for

Report generation with complex table layouts (merged headers, grouped rows)

Template-based document generation with sophisticated table structures

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Merge operations are limited to adjacent cells; non-adjacent merges not supported

Split operations may lose content if merged cell contains text

Complex merge patterns may have unintended visual effects in rendered documents

What makes it unique

Supports both cell merging and splitting with directional control, enabling creation of complex table layouts programmatically.

vs alternatives

More flexible than fixed table structures because it allows dynamic layout adjustments; enables sophisticated report generation with merged headers and grouped data.

table-formatting-with-header-styling

Medium confidence

Solves for

Apply professional formatting to tables for consistent document appearanceStyle table headers for visual distinctionEnable AI agents to create well-formatted tables

Best for

Document generation that requires consistent table formatting

Report generation with professional table styling

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Formatting options are limited to predefined templates; custom formatting requires direct XML manipulation

Formatting is applied at table level; cell-level custom formatting not supported

Color and font choices are limited to document style definitions

What makes it unique

Provides template-based table formatting that applies consistent styling across tables, enabling professional document appearance.

vs alternatives

More convenient than manual cell-by-cell formatting because it applies templates; ensures consistent table appearance across documents.

text-range-formatting-with-multiple-attributes

Medium confidence

Solves for

Best for

Document generation that needs formatted text (emphasis, color coding)

AI agents that apply formatting to highlight important content

Report generation with styled text elements

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Formatting is applied to exact text matches; partial word matches not supported

Font family and size changes may affect document layout; use with caution

Formatting attributes are limited to basic text properties; advanced formatting (shadows, effects) not supported

What makes it unique

Applies multiple formatting attributes simultaneously to text ranges while preserving existing formatting on non-targeted attributes.

vs alternatives

More flexible than paragraph-level formatting because it operates at text range level; enables selective emphasis within paragraphs.

custom-style-creation-and-management

Medium confidence

Solves for

Best for

Enterprise document generation with organizational style standards

Template-based document creation that requires consistent styling

AI agents that apply custom styles to generated content

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Style creation requires detailed property specification; predefined templates not available

Style inheritance is limited to single-level inheritance; complex style hierarchies not supported

Style properties are limited to basic text formatting; advanced properties not supported

What makes it unique

Enables creation of custom styles that are stored in document style sheet and reusable across document, supporting organizational style standards.

vs alternatives

More flexible than predefined styles because custom styles can be tailored to organizational requirements; enables consistent branding across documents.

document-style-listing-and-inspection

Medium confidence

Solves for

Discover available styles in a document for reuseInspect style properties to understand document formatting standardsEnable AI agents to select appropriate styles for content

Best for

Document analysis workflows that need to understand style definitions

Style-aware document generation that reuses existing styles

Documentation of organizational style standards

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Style listing is read-only; does not modify styles

Style properties are extracted as-is; computed properties not calculated

Custom style properties may not be fully captured in listing

What makes it unique

Provides comprehensive style listing with property extraction, enabling style-aware document generation and analysis.

vs alternatives

More informative than simple style enumeration because it includes detailed property information; enables intelligent style selection for generated content.

opc-package-introspection-with-part-enumeration

Medium confidence

Solves for

Analyze document package structure for debugging or advanced customizationDiscover available parts for direct XML manipulationEnable advanced users to understand document internals

Best for

Advanced users performing low-level document customization

Debugging document structure issues

Custom document processing that requires part-level access

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Part enumeration is read-only; does not modify package structure

Part relationships are complex; full relationship graph not always captured

What makes it unique

Provides low-level OPC package introspection for advanced users, enabling direct access to document internals for customization or debugging.

vs alternatives

More detailed than high-level document analysis because it exposes package structure; enables advanced customization not possible through standard tools.

opc-part-xml-and-text-extraction

Medium confidence

Solves for

Extract raw XML from document parts for analysis or custom processingDebug document structure by inspecting underlying XMLEnable advanced users to perform custom XML transformations

Best for

Advanced users performing custom XML transformations

Debugging document structure issues

Custom document processing requiring direct XML access

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

XML extraction is read-only; modifications require separate apply_edit capability

Large XML parts may exceed response size limits

What makes it unique

Provides direct access to OPC part XML content for advanced users, enabling custom XML transformations and low-level document manipulation.

vs alternatives

More powerful than high-level APIs because it exposes raw XML; enables custom transformations not possible through standard tools.

hwpx-to-html-conversion-with-semantic-preservation

Medium confidence

Solves for

Best for

Enterprise knowledge management systems converting legacy Hangeul documents to web-accessible HTML

Accessibility teams requiring semantic HTML output for WCAG compliance

LLM agents that need structured HTML for improved parsing and understanding of document hierarchy

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL

Limitations

Images are omitted by default (image_policy='omitted'); base64 embedding increases payload size significantly

CSS styling is minimal (semantic HTML only); complex formatting requires post-processing with CSS

Form fields and interactive elements in HWPX are converted to static HTML

What makes it unique

vs alternatives

xml-element-search-by-tag-and-attribute

Medium confidence

Solves for

Find specific XML elements for custom processing or analysisLocate elements by attribute values (e.g., find all paragraphs with specific style)Enable advanced users to perform targeted XML queries

Best for

Advanced users performing custom XML queries

Document analysis that requires element-level search

Custom processing based on XML structure

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Search patterns are limited to simple tag/attribute matching; complex XPath not supported

Search results may be large for common tags; filtering required

What makes it unique

Provides element-level XML search by tag and attribute, enabling targeted queries on document structure.

vs alternatives

More precise than full-document XML parsing because it targets specific elements; enables efficient element-level analysis.

edit-planning-with-preview-and-atomic-application

Medium confidence

Solves for

Plan complex document edits with AI assistance before applyingPreview changes to verify correctness before committingEnable safe, reversible document modifications with rollback capability

Best for

Complex document editing workflows requiring change review

AI-driven document modification with human oversight

High-risk document processing requiring change verification

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Preview is text-based; visual preview not available

Rollback requires manual restoration from backup; automatic rollback not supported

What makes it unique

Implements three-step edit workflow (plan → preview → apply) with atomic application, enabling safe document modification with change verification.

vs alternatives

More cautious than direct editing because it requires explicit preview and approval; enables AI-driven editing with human oversight.

document-structure-validation-and-integrity-checking

Medium confidence

Solves for

Verify document integrity after modificationsDetect structural issues that may cause rendering problemsEnable document quality assurance workflows

Best for

Document quality assurance workflows

Post-modification validation to ensure document integrity

Debugging document structure issues

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Validation is structural only; does not check content accuracy

Validation may report false positives for non-standard but valid structures

What makes it unique

Provides comprehensive structure validation with severity levels and correction suggestions, enabling document quality assurance.

vs alternatives

More thorough than simple format checking because it validates structural consistency; enables early detection of issues that may cause rendering problems.

text-convention-linting-for-style-consistency

Medium confidence

Solves for

Enforce text style conventions across documentsDetect inconsistent terminology or formattingEnable document style quality assurance

Best for

Documentation teams enforcing style guidelines

Quality assurance workflows for consistent document style

Automated style checking for organizational standards

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Requires HWPX_MCP_ADVANCED=1 environment variable

Linting rules are predefined; custom rules not easily configurable

Korean language-specific rules may have false positives

What makes it unique

Provides text convention linting with configurable rules for style consistency, enabling automated quality assurance for document style.

vs alternatives

More specific than generic spell-checking because it focuses on style conventions; enables enforcement of organizational writing standards.

mcp-server-transport-abstraction-with-stdio-and-http

Medium confidence

Solves for

Best for

Teams using Claude Desktop or Cursor that prefer stdio transport

Remote document processing workflows requiring HTTP access

Containerized deployments where HTTP is preferred over stdio

Requires

Python >= 3.10

python-hwpx >= 1.9

For HTTP: accessible network port (default 8000)

Limitations

HTTP transport lacks authentication (development-focused); not suitable for untrusted networks

Stdio transport limited to single client connection; HTTP supports multiple concurrent clients

HTTP transport adds ~50-100ms latency per request compared to stdio

What makes it unique

Implements transport abstraction that enables identical tool set to be accessed via stdio or HTTP, supporting diverse deployment scenarios without code duplication.

vs alternatives

More flexible than single-transport servers because it supports both stdio and HTTP; enables use in both desktop (Claude) and server (remote) environments.

hwpx-structure-extraction-to-json-with-toc-generation

Medium confidence

Solves for

Best for

Knowledge management systems building searchable indices of Korean documents

RAG pipelines that need structured document metadata for retrieval ranking

Document analysis tools that classify documents by complexity or structure type

Requires

Python >= 3.10

python-hwpx >= 1.9

Either valid HWPX file path OR base64-encoded HWPX content OR HTTPS URL

Limitations

TOC generation relies on heading style detection; documents with inconsistent heading styles may produce incomplete TOC

Figure extraction captures captions only; image content itself is not extracted (use image-processing capability separately)

Nested table structures are flattened to single-level table list; complex table hierarchies lose nesting information

What makes it unique

vs alternatives

More structured than simple text extraction because it preserves document hierarchy and metadata; more accurate than regex-based heading detection because it uses HWPX style information directly.

keyword-search-with-context-window-retrieval

Medium confidence

Solves for

Best for

Document review workflows where AI needs to locate and analyze specific content

Search-based navigation in large Korean documents without external search engines

Content discovery for RAG systems that need paragraph-level search results

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path

Limitations

Search is linear (O(n) complexity); large documents (>10000 paragraphs) may have noticeable latency

Context window is paragraph-based only; character-level context not supported

Regex patterns require careful escaping; complex regex may have performance impact

What makes it unique

vs alternatives

Faster than full-document loading for targeted searches; more context-aware than simple find-and-replace because it returns surrounding paragraphs for semantic understanding.

search-and-replace-with-split-run-preservation

Medium confidence

Solves for

Best for

Document template automation where placeholders need replacement with formatted content

Bulk document updates (e.g., company name change across multiple documents)

AI-driven document editing workflows that need to preserve formatting

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Split-run handling adds complexity; replacements spanning multiple runs may fail silently if run structure is malformed

Replacement text must be plain text; cannot insert formatted text (use add_paragraph for formatted content)

Case-sensitive matching only; case-insensitive replacement requires pre-processing

What makes it unique

vs alternatives

Preserves formatting metadata that naive string replacement would corrupt; more reliable than regex-based replacement because it understands HWPX XML structure.

paragraph-level-editing-with-insertion-and-deletion

Medium confidence

Solves for

Best for

Document generation workflows that build documents from templates by adding paragraphs

AI agents that need to modify document structure (add sections, delete outdated content)

Automated report generation that assembles documents from content blocks

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Paragraph indices are 0-based and may shift after insertions/deletions; subsequent operations must use updated indices

Style application is limited to predefined styles; custom style creation requires separate capability

Paragraph content must be plain text; formatted text requires separate formatting capability

What makes it unique

vs alternatives

More precise than document-level editing because it operates at paragraph granularity; more flexible than template-based approaches because it supports arbitrary insertion/deletion at any index.

heading-insertion-with-style-application

Medium confidence

Solves for

Best for

Document generation workflows that need to create structured documents with proper heading hierarchy

AI agents that organize content by adding section headers

Automated report generation that creates section structure

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Heading style must exist in document style sheet; custom heading styles require separate style creation

Heading level (1-6) must be valid; invalid levels may default to Heading 1

Heading hierarchy validation is not enforced; can create Heading 3 without Heading 1/2

What makes it unique

Specialized heading insertion that automatically applies correct heading style based on level parameter, ensuring proper document hierarchy and TOC generation.

vs alternatives

More convenient than generic paragraph insertion because it handles style application automatically; ensures proper heading hierarchy for TOC generation.

page-break-insertion-for-document-layout-control

Medium confidence

Solves for

Best for

Document generation workflows that need precise pagination control

Template-based document building where sections must start on new pages

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Page break insertion is position-based; actual page layout depends on content size and margins

Multiple consecutive page breaks may create blank pages

Page break behavior may vary depending on document margins and font sizes

What makes it unique

Provides explicit page break insertion at paragraph level, enabling precise pagination control in programmatically generated documents.

vs alternatives

More reliable than relying on content size for pagination; enables predictable document layout in template-based generation.

memo-annotation-management-add-and-remove

Medium confidence

Solves for

Add review comments or annotations to documents programmaticallyEnable AI agents to mark content for human review with contextual notesRemove outdated annotations or memos from documents

Best for

Document review workflows where AI adds comments for human review

Collaborative editing where AI annotations guide human editors

Requires

Python >= 3.10

python-hwpx >= 1.9

Valid HWPX file path with write permissions

Limitations

Memo text is plain text only; cannot include formatted content

Memo visibility depends on Hangeul application version; may not display in all viewers

Memo indices may shift after paragraph insertions/deletions

What makes it unique

Provides programmatic memo management that integrates with Hangeul application's native annotation system, enabling AI-driven document review workflows.

vs alternatives

More integrated than external annotation systems because memos are stored in document metadata and visible in Hangeul; enables collaborative workflows between AI and human reviewers.

Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.

About

Alternatives to 한글 mcp hwpx MCP Server

Supabase69Platform

Compare →

Tavily MCP Server62MCP Server

AI-optimized web search and content extraction via Tavily MCP.

Compare →

MongoDB MCP Server62MCP Server

Query and manage MongoDB databases and collections via MCP.

Compare →

Firecrawl MCP Server62MCP Server

Scrape websites and extract structured data via Firecrawl MCP.

Compare →

한글 mcp hwpx MCP Server

Capabilities25 decomposed

stateless-hwpx-document-reading-with-context-preservation

hwpx-to-markdown-conversion-with-chunking

document-copying-with-safety-validation

table-creation-with-configurable-dimensions

table-cell-text-modification-with-range-support

table-cell-merging-and-splitting-for-complex-layouts

table-formatting-with-header-styling

text-range-formatting-with-multiple-attributes

custom-style-creation-and-management

document-style-listing-and-inspection

opc-package-introspection-with-part-enumeration

opc-part-xml-and-text-extraction

hwpx-to-html-conversion-with-semantic-preservation

xml-element-search-by-tag-and-attribute

edit-planning-with-preview-and-atomic-application

document-structure-validation-and-integrity-checking

text-convention-linting-for-style-consistency

mcp-server-transport-abstraction-with-stdio-and-http

hwpx-structure-extraction-to-json-with-toc-generation

keyword-search-with-context-window-retrieval

search-and-replace-with-split-run-preservation

paragraph-level-editing-with-insertion-and-deletion

heading-insertion-with-style-application

page-break-insertion-for-document-layout-control

memo-annotation-management-add-and-remove

Related Artifactssharing capabilities

privateGPT

Haystack

llama-parse

Hebbia

Local GPT

rag-memory-epf-mcp

Best For

Known Limitations

Requirements

Input / Output

UnfragileRank

Repository Details

About

Categories

Alternatives to 한글 mcp hwpx MCP Server

Are you the builder of 한글 mcp hwpx MCP Server ?

Get the weekly brief

Data Sources

한글 mcp hwpx MCP Server

Capabilities25 decomposed

stateless-hwpx-document-reading-with-context-preservation

hwpx-to-markdown-conversion-with-chunking

document-copying-with-safety-validation

table-creation-with-configurable-dimensions

table-cell-text-modification-with-range-support

table-cell-merging-and-splitting-for-complex-layouts

table-formatting-with-header-styling

text-range-formatting-with-multiple-attributes

custom-style-creation-and-management

document-style-listing-and-inspection

opc-package-introspection-with-part-enumeration

opc-part-xml-and-text-extraction

hwpx-to-html-conversion-with-semantic-preservation

xml-element-search-by-tag-and-attribute

edit-planning-with-preview-and-atomic-application

document-structure-validation-and-integrity-checking

text-convention-linting-for-style-consistency

mcp-server-transport-abstraction-with-stdio-and-http

hwpx-structure-extraction-to-json-with-toc-generation

keyword-search-with-context-window-retrieval

search-and-replace-with-split-run-preservation

paragraph-level-editing-with-insertion-and-deletion

heading-insertion-with-style-application

page-break-insertion-for-document-layout-control

memo-annotation-management-add-and-remove

Related Artifactssharing capabilities

privateGPT

Haystack

llama-parse

Hebbia

Local GPT

rag-memory-epf-mcp