Skip to main content
Glama
anton-prosterity

Documentation Search MCP Server

CHANGELOG.md6.24 kB
# Changelog All notable changes to Documentation Search Enhanced MCP Server will be documented in this file. ## [Unreleased] ## [1.5.0] - 2025-12-17 ### Added - GitHub Release assets `docs_site_index.json` and `docs_site_index.json.gz` for Serper-free docs search. - Auto-download support for the prebuilt index (env: `DOCS_SITE_INDEX_AUTO_DOWNLOAD`, `DOCS_SITE_INDEX_MAX_AGE_HOURS`, `DOCS_SITE_INDEX_URL(S)`, `DOCS_SITE_INDEX_PATH`). - CI workflow to build and upload the index on release publish. - HTML link discovery fallback for docs sites without sitemaps/indexes (e.g. `react.dev`). ### Changed - `SERPER_API_KEY` is now optional; when unset, documentation search uses a prebuilt docs index (auto-downloaded from GitHub Releases on startup) plus on-site docs search (MkDocs/Sphinx indexes when available). - Serper-free sitemap search now prefers live sitemap refreshes when available, falling back to the prebuilt index (and HTML link discovery for sites without sitemaps, e.g. `react.dev`). - Playwright scraping now falls back to `httpx` when Playwright cannot launch (common in sandboxed environments). - Improved Codex CLI reliability: avoid stdio handshake stalls and reduce default startup output noise. ## [1.4.1] - 2025-12-16 ### Changed - Simplified version_resolver.py from 376 to 190 lines by consolidating duplicate subprocess methods - Replaced 1000+ lines of hardcoded learning paths with dynamic 30-line generator - Reduced main.py from 2,923 to 1,866 lines (removed 1,057 lines of AI-generated bloat) - Removed excessive docstrings and commented placeholder code - Improved code maintainability and readability ## [1.4.0] - 2025-12-16 ### Added - Codex CLI setup instructions in `README.md` and `samples/README.md`. ### Changed - MCP server name now reports as `documentation_search_enhanced` (was `docs`) for clearer tool namespacing in MCP clients. - `get_docs` and `semantic_search` now publish a simpler `libraries` input schema (array of strings) while still accepting a single/comma-separated string. ### Fixed - All internal logging/diagnostic `print()` output now goes to stderr to avoid corrupting stdio MCP transport (important for Codex CLI). ## [1.3.2] - 2025-09-24 ### Added - AGENTS.md contributor guide tailored to this repository. ### Changed - README: clearer install (`uv sync --all-extras --all-groups`) and testing instructions (core vs. example tests). - Example project: added `httpx` dependency and documented `python -m pytest` usage. - Codebase formatted with Black; no functional changes. ## [1.3.1] - 2025-06-28 ### Fixed - Republishing to ensure latest features are correctly uploaded to PyPI. ## [1.3.0] - 2025-06-26 ### Added - **Project-Aware Security Scan**: New `scan_project_dependencies` tool to scan all dependencies in a project (`pyproject.toml`, `requirements.txt`, etc.) for vulnerabilities. - **Project Boilerplate Generation**: New `generate_project_starter` tool to create starter projects from templates (FastAPI, React+Vite). - **Multi-Library Search**: Enhanced `get_docs` and `semantic_search` to accept a list of libraries, allowing for comparison and discovery across multiple documentation sources. - **Local Dev Environment Management**: New `manage_dev_environment` tool to generate `docker-compose.yml` files for services like `postgres`, `redis`, and `rabbitmq`. ### Changed - Pivoted away from the complex Test Execution feature in favor of the more robust Docker Environment Management. ### Fixed - Resolved numerous bugs in the multi-library search and project generation features through iterative testing and debugging. - Corrected a `ModuleNotFoundError` by adding `PyYAML` as a dependency. ## [1.2.0] - 2025-06-26 ### Added - **Production Essential Tools** (5 new libraries): - Prometheus - Monitoring and metrics - Grafana - Data visualization and dashboards - Elasticsearch - Search and analytics engine - Celery - Distributed task queue - ESLint - JavaScript linting - **Enhanced Learning Paths**: - Multi-library learning paths that intelligently use different documentation sources - New comprehensive paths: frontend-development, backend-development, fullstack-development, devops - Individual library paths now reference multiple related libraries for better learning - **Advanced Search Features**: - `semantic_search` - AI-powered search with relevance scoring - `filtered_search` - Filter by content type, difficulty level, code examples - `get_code_examples` - Extract curated code examples from documentation - **Environment-Aware Configuration**: - Support for development, testing, staging, and production environments - Environment-specific settings for caching, logging, and rate limiting - Auto-detection from ENVIRONMENT/ENV variables - **Enhanced Features**: - Smart query expansion for better search results - Content difficulty assessment - Cross-reference detection - Contextual recommendations - Structured logging with performance tracking ### Changed - Learning paths now use multiple libraries for comprehensive education - Configuration structure enhanced with server_config, features, and priorities - Cache TTL now environment-aware (1 hour dev, 24 hours production) ### Fixed - Removed duplicate main.py file that was causing RuntimeWarning - Fixed linter errors in security tool implementations ## [1.1.0] - 2025-06-26 ### Added - OSINT vulnerability scanning using OSV Database, GitHub Security Advisories, and Safety DB - Security scoring system (0-100 scale) for libraries - New security-focused tools: - `scan_library_vulnerabilities` - Full vulnerability scan - `get_security_summary` - Quick security overview - `compare_library_security` - Compare multiple libraries - `suggest_secure_libraries` - Enhanced suggestions with security scores - Support for multiple ecosystems (PyPI, npm, Maven, Go) - Auto-approve configuration for sensitive libraries ## [1.0.0] - 2025-06-25 ### Initial Release - Support for 99 popular development libraries - Basic documentation search and retrieval - Library suggestions with fuzzy matching - Health check for documentation sources - Simple caching system - AWS MCP-style deployment via uvx

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/anton-prosterity/documentation-search-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server