oarx-docs-mcp
MCP server for semantic search across the complete ObjectARX 2026 documentation set. Indexes 89,715 text chunks from 22 guides in ChromaDB with hybrid BM25 + dense retrieval.
README
oarx-docs-mcp
MCP server for semantic search across the complete ObjectARX 2026 documentation set. 89,715 text chunks from 22 guides indexed in ChromaDB with hybrid BM25 + dense retrieval.
Overview
oarx-docs-mcp indexes 89,715 text chunks from 22 documentation guides into a local ChromaDB vector database and exposes them through the Model Context Protocol (MCP) for use with Claude Code, Claude Desktop, or any MCP-compatible client.
Highlights
- 3 MCP tools —
search_docs,lookup_class,list_guides - 89,715 indexed chunks from 22 documentation guides (ObjectARX SDK + Autodesk CloudHelp)
- Hybrid retrieval — BM25 sparse + dense vectors + Reciprocal Rank Fusion
- Cross-encoder reranking with BAAI/bge-reranker-v2-m3
- Query optimization — AutoLISP detection, C++ API expansion, guide-specific boosting
- Container-ready — Containerfile + compose.yaml (Podman/Docker)
- Claude Code plugin — 4 skills + 2 slash commands for ObjectARX development
Data Sources
| Source | Guides | Pages | Description |
|---|---|---|---|
| ObjectARX SDK (CHM) | 8 | ~45,800 | C++ Reference, .NET Reference, Developer Guides, Migration, Interop |
| Autodesk CloudHelp (web) | 14 | ~5,000 | AutoLISP, ActiveX/VBA, JavaScript, DXF, Customization |
MCP Tools
| Tool | Description |
|---|---|
list_guides |
List all 22 available documentation guides |
search_docs |
Semantic search with optional guide filter and result limit |
lookup_class |
Look up a class or function by name (title match with semantic fallback) |
Quick Start
# Clone and install
cd mcp/oarx-docs-mcp
uv sync
# Download web documentation (~5,000 pages)
OARX_DATA_DIR=./data uv run python -m oarx_docs_mcp.scraper
# Index CHM files from ObjectARX SDK (optional, requires 7-Zip)
OARX_DATA_DIR=./data OARX_CHM_DIR=/path/to/sdk/docs uv run python -m oarx_docs_mcp.indexer
# Or index web docs only (no SDK needed)
OARX_DATA_DIR=./data uv run python -m oarx_docs_mcp.indexer
Client Configuration
<details> <summary><b>Claude Code</b></summary>
Add to ~/.claude.json (or project-level .claude.json):
{
"mcpServers": {
"oarx-docs": {
"command": "uv",
"args": [
"run",
"--directory", "/path/to/oarx-docs-mcp",
"python", "-m", "oarx_docs_mcp"
],
"env": {
"OARX_DATA_DIR": "/path/to/oarx-docs-mcp/data"
}
}
}
}
<details> <summary>Windows example</summary>
{
"mcpServers": {
"oarx-docs": {
"command": "uv",
"args": [
"run",
"--directory", "C:/Users/user/oarx-docs-mcp",
"python", "-m", "oarx_docs_mcp"
],
"env": {
"OARX_DATA_DIR": "C:/Users/user/oarx-docs-mcp/data"
}
}
}
}
</details>
</details>
<details> <summary><b>Claude Desktop</b></summary>
Add to Claude Desktop's MCP configuration file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"oarx-docs": {
"command": "uv",
"args": [
"run",
"--directory", "/path/to/oarx-docs-mcp",
"python", "-m", "oarx_docs_mcp"
],
"env": {
"OARX_DATA_DIR": "/path/to/oarx-docs-mcp/data"
}
}
}
}
Restart Claude Desktop to pick up the new configuration.
</details>
<details> <summary><b>OpenAI Codex CLI</b></summary>
Add to ~/.codex/config.toml (global) or .codex/config.toml (project):
[mcp_servers.oarx-docs]
command = "uv"
args = ["run", "--directory", "/path/to/oarx-docs-mcp", "python", "-m", "oarx_docs_mcp"]
[mcp_servers.oarx-docs.env]
OARX_DATA_DIR = "/path/to/oarx-docs-mcp/data"
Or via CLI:
codex mcp add oarx-docs --env OARX_DATA_DIR=/path/to/data -- uv run --directory /path/to/oarx-docs-mcp python -m oarx_docs_mcp
</details>
Container Setup
# Build image
podman-compose build
# Download web docs
podman-compose --profile scraper run --rm scraper
# Index (requires OARX_CHM_DIR env var for CHM files)
OARX_CHM_DIR=/path/to/sdk/docs podman-compose --profile indexer run --rm indexer
# Run server
podman-compose up server
Configuration
All settings are configurable via environment variables:
| Variable | Default | Description |
|---|---|---|
OARX_CHM_DIR |
/docs |
Directory containing .chm files from ObjectARX SDK |
OARX_DATA_DIR |
/app/data |
Data directory for ChromaDB and extracted HTML |
OARX_EMBEDDING_MODEL |
BAAI/bge-m3 |
Sentence-transformers model name |
OARX_CHUNK_SIZE |
800 |
Text chunk size in characters |
OARX_CHUNK_OVERLAP |
200 |
Overlap between adjacent chunks |
OARX_7Z_PATH |
auto-detect | Path to 7z executable |
OARX_DB_SOURCE |
github |
Database source: github, huggingface, or local |
OARX_DB_VERSION |
0.1.0 |
Database version to download |
OARX_GITHUB_REPO |
dantte-lp/oarx-docs-mcp |
GitHub repository for release assets |
Development
# Install with dev dependencies
uv sync --all-extras
# Run tests
uv run python -m pytest tests/ -v
# Lint
uv run ruff check src/ tests/
# Type check
uv run mypy
Tech Stack
- Python 3.13+ with uv
- MCP SDK (FastMCP, stdio transport)
- ChromaDB (persistent vector store, HNSW + cosine)
- sentence-transformers (
all-MiniLM-L6-v2, 384-dim embeddings) - BeautifulSoup4 + lxml (HTML parsing)
- httpx (web documentation scraping)
- Podman / podman-compose (containerization)
Project Structure
oarx-docs-mcp/
├── src/oarx_docs_mcp/
│ ├── __main__.py # Entry point
│ ├── config.py # Settings + guide registries
│ ├── data_manager.py # Database auto-download
│ ├── indexer.py # CHM/HTML → ChromaDB
│ ├── retriever.py # BM25 + dense + RRF + reranking
│ ├── scraper.py # CloudHelp downloader
│ └── server.py # MCP tools (3 tools)
├── tests/ # pytest + pytest-asyncio
├── plugin/ # Claude Code plugin
├── docs/
│ ├── en/ # English documentation
│ └── ru/ # Russian documentation
├── Containerfile # OCI container image
└── compose.yaml # Podman/Docker services
Documentation
| Document | EN | RU |
|---|---|---|
| Architecture | architecture.md | architecture.md |
| Installation | installation.md | installation.md |
| Configuration | configuration.md | configuration.md |
| Usage | usage.md | usage.md |
| Troubleshooting | troubleshooting.md | troubleshooting.md |
| Changelog | CHANGELOG.md | — |
| Contributing | CONTRIBUTING.md | CONTRIBUTING.RU.md |
Contributing
See CONTRIBUTING.md (RU).
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.