Orionbelt Analytics

v0.5.0Analyticsstable

Ontology-based MCP server that analyzes database schemas (PostgreSQL, Snowflake, ClickHouse, Dremio) and generates RDF/OWL ontologies with SQL mappings for fan-trap-free Text-to-SQL.

agenticagentic-aiai-analyticsclaude-desktopclickhouse
Share:
29
Stars
0
Downloads
0
Weekly
0/5

What is Orionbelt Analytics?

Orionbelt Analytics is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to ontology-based mcp server that analyzes database schemas (postgresql, snowflake, clickhouse, dremio) and generates rdf/owl ontologies with sql mappings for fan-trap-free text-to-sql.

Ontology-based MCP server that analyzes database schemas (PostgreSQL, Snowflake, ClickHouse, Dremio) and generates RDF/OWL ontologies with SQL mappings for fan-trap-free Text-to-SQL.

This server falls under the Analytics category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Ontology-based MCP server that analyzes database schemas (Po

Use Cases

Analyze database schemas
Generate RDF ontologies
Text-to-SQL optimization
ralfbecher

Maintainer

LicenseNOASSERTION
Languagepython
Versionv0.5.0
UpdatedMay 11, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

PIP

pip install orionbelt-analytics

Manual Installation

pip install orionbelt-analytics

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Orionbelt Analytics

OrionBelt Analytics is an ontology-based MCP server that connects to relational databases — PostgreSQL, MySQL, Snowflake, ClickHouse, Dremio, BigQuery, DuckDB/MotherDuck, and Databricks SQL — analyzes their schemas, and generates RDF/OWL ontologies with embedded SQL mappings to enable accurate, fan-trap-free Text-to-SQL. Its OBQC (Ontology-Based Query Check) layer deterministically validates every generated SQL statement against the ontology before execution, catching table/column errors, invalid joins, aggregation mistakes, and fan-trap conditions that LLMs alone would silently pass. Data analysts and BI teams use it to get reliable natural-language SQL generation over complex multi-table schemas without getting incorrect numbers from fan-trap joins.

Prerequisites

  • Python 3.13 or later installed
  • uv package manager installed (curl -LsSf https://astral.sh/uv/install.sh | sh)
  • Access credentials for at least one supported database (PostgreSQL, Snowflake, ClickHouse, BigQuery, etc.)
  • An MCP-compatible client such as Claude Desktop or Claude Code
  • Optional: ChromaDB and Oxigraph are bundled; Qdrant is not required for core functionality
1

Clone the repository and install dependencies

Clone the OrionBelt Analytics repository and use uv to sync all Python dependencies. Python 3.13+ is required.

git clone https://github.com/ralfbecher/orionbelt-analytics
cd orionbelt-analytics
uv sync
2

Create and configure the .env file

Copy the environment template and edit it with your database credentials. Set at minimum the variables for one database backend (for example POSTGRES_HOST, POSTGRES_PORT, POSTGRES_DATABASE, POSTGRES_USERNAME, POSTGRES_PASSWORD).

cp .env.template .env
3

Start the OrionBelt Analytics server

The server starts on HTTP transport at localhost:9000 by default. The port is configurable via MCP_SERVER_PORT in .env.

uv run server.py
4

Connect Claude Desktop via mcp-remote

Because OrionBelt uses HTTP transport, Claude Desktop connects via the mcp-remote proxy. Add the following to your claude_desktop_config.json.

5

Connect and discover your schema

Once connected in Claude Desktop, use the connect_database tool to establish a connection, then discover_schema to analyze the schema, build the GraphRAG index, and generate the initial ontology.

6

Generate the RDF/OWL ontology

After discovery, call generate_ontology to produce a fully annotated Turtle (.ttl) file with oba: namespace SQL mappings. This ontology powers OBQC validation for all subsequent SQL generation.

Orionbelt Analytics Examples

Client configuration

OrionBelt Analytics uses HTTP transport. Connect Claude Desktop via mcp-remote. Start the server first, then add this to claude_desktop_config.json.

{
  "mcpServers": {
    "OrionBelt-Analytics": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:9000/mcp",
        "--transport",
        "http-only"
      ]
    }
  }
}

Prompts to try

These prompts walk through the typical workflow: connect to a database, discover its schema, generate an ontology, and then run validated Text-to-SQL queries.

- "Connect to my PostgreSQL database and discover the schema for the 'sales' schema"
- "Generate an RDF/OWL ontology from the connected schema"
- "How many orders were placed per customer last month? Generate and run the SQL"
- "Find all join paths between the orders and products tables"
- "Generate a bar chart of monthly revenue for 2024 using the orders table"

Troubleshooting Orionbelt Analytics

connect_database fails with connection timeout or authentication error

Verify all database credentials in .env match exactly (host, port, database name, username, password). For cloud databases like Snowflake or BigQuery, also check that the account identifier and warehouse/project settings are correct.

execute_sql_query is blocked by OBQC with 'fan-trap detected' error

This is OBQC working correctly — the generated SQL has a join across multiple one-to-many relationships that would multiply results. Rephrase the question to compute aggregations separately per relationship or ask OrionBelt to rewrite the query using subqueries.

Claude Desktop cannot reach the server with 'connection refused' on localhost:9000

Ensure the server is running (uv run server.py) before starting Claude Desktop. If the port is in use, set MCP_SERVER_PORT to a free port in .env and update the mcp-remote URL in claude_desktop_config.json accordingly.

Frequently Asked Questions about Orionbelt Analytics

What is Orionbelt Analytics?

Orionbelt Analytics is a Model Context Protocol (MCP) server that ontology-based mcp server that analyzes database schemas (postgresql, snowflake, clickhouse, dremio) and generates rdf/owl ontologies with sql mappings for fan-trap-free text-to-sql. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Orionbelt Analytics?

Install via pip with: pip install orionbelt-analytics. Then configure your AI client to connect to this MCP server.

Which AI clients work with Orionbelt Analytics?

Orionbelt Analytics works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Orionbelt Analytics free to use?

Yes, Orionbelt Analytics is open source and available under the NOASSERTION license. You can use it freely in both personal and commercial projects.

Browse More Analytics MCP Servers

Explore all analytics servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "orionbelt-analytics": { "command": "pip", "args": ["install", "orionbelt-analytics"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Orionbelt Analytics?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides