OpenZIM MCP Server

🧠 Built for LLM Intelligence

OpenZIM MCP transforms static ZIM archives into dynamic knowledge engines for Large Language Models. Unlike basic file readers, this tool provides intelligent, structured access that LLMs need to effectively navigate and understand vast knowledge repositories.

🚀 Why LLMs Love OpenZIM MCP:

• Smart Navigation: Browse by namespace (articles, metadata, media) instead of blind searching

• Context-Aware Discovery: Get article structure, relationships, and metadata for deeper understanding

• Intelligent Search: Advanced filtering, auto-complete suggestions, and relevance-ranked results

• Performance Optimized: Cached operations and pagination prevent timeouts on massive archives

• Relationship Mapping: Extract internal/external links to understand content connections

Whether you're building a research assistant, knowledge chatbot, or content analysis system, OpenZIM MCP gives your LLM the structured access patterns it needs to unlock the full potential of offline knowledge archives. No more fumbling through raw text dumps! 🎯

OpenZIM MCP is a modern, secure, and high-performance MCP (Model Context Protocol) server that enables AI models to access and search ZIM format knowledge bases offline.

ZIM (Zeno IMproved) is an open file format developed by the openZIM project, designed specifically for offline storage and access to website content. The format supports high compression rates using Zstandard compression (default since 2021) and enables fast full-text searching, making it ideal for storing entire Wikipedia content and other large reference materials in relatively compact files. The openZIM project is sponsored by Wikimedia CH and supported by the Wikimedia Foundation, ensuring the format's continued development and adoption for offline knowledge access, especially in environments without reliable internet connectivity.

✨ Features

• 🔒 Security First: Comprehensive input validation and path traversal protection

• ⚡ High Performance: Intelligent caching and optimized ZIM file operations

• 🧠 Smart Retrieval: Automatic fallback from direct access to search-based retrieval for reliable entry access

• 🧪 Well Tested: 90%+ test coverage with comprehensive test suite

• 🏗️ Modern Architecture: Modular design with dependency injection

• 📝 Type Safe: Full type annotations throughout the codebase

• 🔧 Configurable: Flexible configuration with validation

• 📊 Observable: Structured logging and health monitoring

🚀 Quick Start

Installation

Development Installation

For contributors and developers:

Prepare ZIM Files

Download ZIM files (e.g., Wikipedia, Wiktionary, etc.) from the Kiwix Library and place them in a directory:

Running the Server

MCP Configuration

Add to your MCP client configuration:

Alternative configuration using Python module:

For development (from source):

🛠️ Development

Running Tests

ZIM Test Data Integration

OpenZIM MCP integrates with the official zim-testing-suite for comprehensive testing with real ZIM files:

The test data includes:

• Basic files: Small ZIM files for essential testing

• Real content: Actual Wikipedia/Wikibooks content for integration testing

• Invalid files: Malformed ZIM files for error handling testing

• Special cases: Embedded content, split files, and edge cases

Test files are automatically organized by category and priority level.

Code Quality

Project Structure

📚 API Reference

Available Tools

list_zim_files - List all ZIM files in allowed directories

No parameters required.

search_zim_file - Search within ZIM file content

Required parameters:

• zim_file_path (string): Path to the ZIM file

• query (string): Search query term

Optional parameters:

• limit (integer, default: 10): Maximum number of results to return

• offset (integer, default: 0): Starting offset for results (for pagination)

get_zim_entry - Get detailed content of a specific entry in a ZIM file

Required parameters:

• zim_file_path (string): Path to the ZIM file

• entry_path (string): Entry path, e.g., 'A/Some_Article'

Optional parameters:

• max_content_length (integer, default: 100000, minimum: 1000): Maximum length of returned content

Smart Retrieval Features:

• Automatic Fallback: If direct path access fails, automatically searches for the entry and uses the exact path found

• Path Mapping Cache: Caches successful path mappings for improved performance on repeated access

• Enhanced Error Guidance: Provides clear guidance when entries cannot be found, suggesting alternative approaches

• Transparent Operation: Works seamlessly regardless of path encoding differences (spaces vs underscores, URL encoding, etc.)

get_zim_metadata - Get ZIM file metadata from M namespace entries

Required parameters:

• zim_file_path (string): Path to the ZIM file

Returns: JSON string containing ZIM metadata including entry counts, archive information, and metadata entries like title, description, language, creator, etc.

get_main_page - Get the main page entry from W namespace

Required parameters:

• zim_file_path (string): Path to the ZIM file

Returns: Main page content or information about the main page entry.

list_namespaces - List available namespaces and their entry counts

Required parameters:

• zim_file_path (string): Path to the ZIM file

Returns: JSON string containing namespace information with entry counts, descriptions, and sample entries for each namespace (C, M, W, X, etc.).

browse_namespace - Browse entries in a specific namespace with pagination

Required parameters:

• zim_file_path (string): Path to the ZIM file

• namespace (string): Namespace to browse (C, M, W, X, A, I, etc.)

Optional parameters:

• limit (integer, default: 50, range: 1-200): Maximum number of entries to return

• offset (integer, default: 0): Starting offset for pagination

Returns: JSON string containing namespace entries with titles, content previews, and pagination information.

search_with_filters - Search within ZIM file content with advanced filters

Required parameters:

• zim_file_path (string): Path to the ZIM file

• query (string): Search query term

Optional parameters:

• namespace (string): Optional namespace filter (C, M, W, X, etc.)

• content_type (string): Optional content type filter (text/html, text/plain, etc.)

• limit (integer, default: 10, range: 1-100): Maximum number of results to return

• offset (integer, default: 0): Starting offset for pagination

Returns: Filtered search results with namespace and content type information.

get_search_suggestions - Get search suggestions and auto-complete

Required parameters:

• zim_file_path (string): Path to the ZIM file

• partial_query (string): Partial search query (minimum 2 characters)

Optional parameters:

• limit (integer, default: 10, range: 1-50): Maximum number of suggestions to return

Returns: JSON string containing search suggestions based on article titles and content.

get_article_structure - Extract article structure and metadata

Required parameters:

• zim_file_path (string): Path to the ZIM file

• entry_path (string): Entry path, e.g., 'C/Some_Article'

Returns: JSON string containing article structure including headings, sections, metadata, and word count.

extract_article_links - Extract internal and external links from an article

Required parameters:

• zim_file_path (string): Path to the ZIM file

• entry_path (string): Entry path, e.g., 'C/Some_Article'

Returns: JSON string containing categorized links (internal, external, media) with titles and metadata.

Examples

Listing ZIM files

Response:

Searching ZIM files

Response:

Getting ZIM entries

Response:

Smart Retrieval in Action

Example: Automatic path resolution

Response (showing smart retrieval working):

get_server_health - Get server health and statistics

No parameters required.

Returns:

• Server status and performance metrics

• Cache statistics

• Configuration information

• Instance tracking information

• Conflict detection results

Example Response:

get_server_configuration - Get detailed server configuration

No parameters required.

Returns: Comprehensive server configuration including diagnostics, validation results, and conflict detection.

Example Response:

diagnose_server_state - Comprehensive server diagnostics

No parameters required.

Returns: Detailed diagnostic information including instance conflicts, configuration validation, file accessibility checks, and actionable recommendations.

Example Response:

resolve_server_conflicts - Identify and resolve server conflicts

No parameters required.

Returns: Results of conflict resolution including cleanup actions and recommendations.

Example Response:

Additional Search Examples

Computer-related search:

Response:

Getting detailed content:

Response:

🎯 Advanced Knowledge Retrieval Examples

Getting ZIM metadata:

Response:

Browsing a namespace:

Response:

Filtered search:

Getting article structure:

Response:

Getting search suggestions:

Response:

🔧 Server Management and Diagnostics Examples

Getting server health:

Response:

Diagnosing server state:

Response:

Resolving server conflicts:

Response:

🎯 ZIM Entry Retrieval Best Practices

Smart Retrieval System

OpenZIM MCP implements an intelligent entry retrieval system that automatically handles path encoding inconsistencies common in ZIM files:

How It Works:

1. Direct Access First: Attempts to retrieve the entry using the provided path exactly as given

2. Automatic Fallback: If direct access fails, automatically searches for the entry using various search terms

3. Path Mapping Cache: Caches successful path mappings to improve performance for repeated access

4. Enhanced Error Guidance: Provides clear guidance when entries cannot be found

Benefits for LLM Users:

• Transparent Operation: No need to understand ZIM path encoding complexities

• Single Tool Call: Eliminates the need for manual search-first methodology

• Reliable Results: Consistent success across different path formats (spaces vs underscores, URL encoding, etc.)

• Performance Optimized: Cached mappings improve repeated access speed

Example Scenarios Handled Automatically:

• A/Test Article → A/Test_Article (space to underscore conversion)

• C/Café → C/Caf%C3%A9 (URL encoding differences)

• A/Some-Page → A/Some_Page (hyphen to underscore conversion)

Usage Recommendations

For Direct Entry Access:

When Entry Not Found: The system will automatically provide guidance:

⚠️ Important Notes and Limitations

Content Length Requirements

• The max_content_length parameter for get_zim_entry must be at least 1000 characters

• Content longer than the specified limit will be truncated with a note showing the total character count

Search Behavior

• Search results may include articles that contain the search terms in various contexts

• Results are ranked by relevance but may not always be directly related to the primary meaning of the search term

• Search snippets provide a preview of the content but may not show the exact location where the search term appears

File Format Support

• Currently supports ZIM files (Zeno IMproved format)

• Tested with Wikipedia ZIM files (e.g., wikipedia_en_100_2025-08.zim)

• File paths must be properly escaped in JSON (use \\ for Windows paths)

🔄 Multi-Server Instance Management

OpenZIM MCP includes advanced multi-server instance tracking and conflict detection to ensure reliable operation when multiple server instances are running.

Instance Tracking Features

• Automatic Instance Registration: Each server instance is automatically registered with a unique process ID and configuration hash

• Conflict Detection: Detects when multiple servers with different configurations are accessing the same directories

• Stale Instance Cleanup: Automatically identifies and cleans up orphaned instance files from terminated processes

• Configuration Validation: Ensures all server instances use compatible configurations

Conflict Types

1. Configuration Mismatch: Multiple servers with different settings accessing the same directories

2. Multiple Instances: Multiple servers running simultaneously (may cause confusion)

3. Stale Instances: Orphaned instance files from terminated processes

Automatic Conflict Warnings

OpenZIM MCP automatically includes conflict warnings in search results and file listings when issues are detected:

Best Practices

• Use diagnose_server_state() regularly to check for conflicts

• Run resolve_server_conflicts() to clean up stale instances

• Ensure all server instances use the same configuration when accessing shared directories

• Monitor server health with get_server_health() for instance tracking information

🔧 Configuration

OpenZIM MCP supports configuration through environment variables with the OPENZIM_MCP_ prefix:

Configuration Options

🔒 Security Features

• Path Traversal Protection: Secure path validation prevents access outside allowed directories

• Input Sanitization: All user inputs are validated and sanitized

• Resource Management: Proper cleanup of ZIM archive resources

• Error Handling: Sanitized error messages prevent information disclosure

• Type Safety: Full type annotations prevent type-related vulnerabilities

🚀 Performance Features

• Intelligent Caching: LRU cache with TTL for frequently accessed content

• Resource Pooling: Efficient ZIM archive management

• Optimized Content Processing: Fast HTML to text conversion

• Lazy Loading: Components initialized only when needed

• Memory Management: Proper cleanup and resource management

🧪 Testing

The project includes comprehensive testing with 90%+ coverage using both mock data and real ZIM files:

Test Categories

• Unit Tests: Individual component testing with mocks

• Integration Tests: End-to-end functionality testing with real ZIM files

• Security Tests: Path traversal and input validation testing

• Performance Tests: Cache and resource management testing

• Format Compatibility: Testing with various ZIM file formats and versions

• Error Handling: Testing with invalid and malformed ZIM files

Test Infrastructure

OpenZIM MCP uses a hybrid testing approach:

1. Mock-based tests: Fast unit tests using mocked libzim components

2. Real ZIM file tests: Integration tests using official zim-testing-suite files

3. Automatic test data management: Download and organize test files as needed

Test Data Sources

• Built-in test data: Basic test files included in the repository

• zim-testing-suite integration: Official test files from the OpenZIM project

• Environment variable support: ZIM_TEST_DATA_DIR for custom test data locations

Test Markers

Tests are organized with pytest markers:

• @pytest.mark.requires_zim_data: Tests requiring ZIM test data files

• @pytest.mark.integration: Integration tests

• @pytest.mark.slow: Long-running tests

📈 Monitoring

OpenZIM MCP provides built-in monitoring capabilities:

• Health Checks: Server health and status monitoring

• Cache Metrics: Cache hit rates and performance statistics

• Structured Logging: JSON-formatted logs for easy parsing

• Error Tracking: Comprehensive error logging and tracking

🔄 Versioning

This project uses Semantic Versioning with automated version management through release-please.

Automated Releases

Version bumps and releases are automated based on Conventional Commits:

• feat: - New features (minor version bump)

• fix: - Bug fixes (patch version bump)

• feat!: or BREAKING CHANGE: - Breaking changes (major version bump)

• perf: - Performance improvements (patch version bump)

• docs:, style:, refactor:, test:, chore: - No version bump

Release Process

The project uses an improved, consolidated release system with automatic validation:

1. Automatic (Recommended): Push conventional commits → Release Please creates PR → Merge PR → Automatic release

2. Manual: Use GitHub Actions UI for direct control over releases

3. Emergency: Push tags directly for critical fixes

Key Features:

• ✅ Zero-touch releases from main branch

• ✅ Automatic version synchronization validation

• ✅ Comprehensive testing before every release

• ✅ Improved error handling and rollback capabilities

• ✅ Branch protection prevents broken releases

For detailed instructions, see Release Process Guide.

Commit Message Format

Examples:

🤝 Contributing

1. Fork the repository

2. Create a feature branch (git checkout -b feature/amazing-feature)

3. Make your changes

4. Run tests (make check)

5. Use conventional commit messages (git commit -m 'feat: add amazing feature')

6. Push to the branch (git push origin feature/amazing-feature)

7. Open a Pull Request

Development Guidelines

• Follow PEP 8 style guidelines

• Add type hints to all functions

• Write tests for new functionality

• Update documentation as needed

• Use conventional commit messages for automatic versioning

• Ensure all tests pass before submitting

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Kiwix for the ZIM format and libzim library

• MCP for the Model Context Protocol

• The open-source community for the excellent libraries used in this project