MCP Connector

Interact with TigerGraph Databases via Natural Language

Community MCP server for TigerGraph - schema introspection, GSQL query execution, vertex/edge CRUD, and admin monitoring tools.

Works with tigergraph

90
Spark score
out of 100
Updated 5 months ago
Version 1.0.0
Models
universal

Add to Favorites

Why it matters

Leverage natural language to interact with your TigerGraph databases. This asset provides a Python MCP server that exposes database operations as structured tools for seamless integration with LLM-based chat interfaces.

Outcomes

What it gets done

01

Execute GSQL queries and manage graph schemas using natural language commands.

02

Programmatically create, alter, and upsert vertices and edges, including vector attributes.

03

Introspect graph schemas, list available queries, and retrieve UDF/algorithm catalogs.

04

Monitor database service status, component versions, and resource utilization (admin functions).

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/vb-tigergraph | bash

Capabilities

Tools your agent gets

schema_introspection

Retrieve the complete graph schema including vertex and edge types

query_view

Get a list of available queries that can be executed

query_execution

Execute installed GSQL queries or raw GSQL strings with parameters

create_vertex_edge

Programmatically create vertices and edges in the graph

alter_vertex

Modify a vertex to enable vector attribute support

upsert_vertex_edge

Update vertex and edge attributes including vector attributes

udf_algorithm_listing

Retrieve installed user-defined functions and GDS algorithm catalogs

displayService_Status

Display database service status showing which services are online

+4 tools

Overview

TigerGraph MCP Server

A community MCP server exposing TigerGraph schema introspection, GSQL query execution, vertex/edge CRUD with vector attributes, and superuser-gated admin monitoring. Use when an agent needs natural-language access to a TigerGraph database's schema, queries, or graph data, given TigerGraph 4.2.0+ and Python 3.12+.

What it does

TigerGraph MCP Server is a community Python MCP server exposing TigerGraph graph-database operations as structured tools: schema introspection, GSQL query listing and execution, vertex/edge creation and updates (including vector attributes), user-defined function and GDS algorithm catalog listing, and administrative service monitoring. It also ships a built-in chat-bot with session management that automatically handles TigerGraph's SECRET and TOKEN retrieval, and supports both Anthropic Claude and OpenAI models for that chat interface.

When to use - and when NOT to

Use this connector when an agent needs to query or modify a TigerGraph graph database through natural language - inspecting the schema, running installed GSQL queries or raw GSQL strings with parameters, creating or updating vertices and edges, or listing available algorithms. Administrative tools (displayService_Status, displayComponent_Version, displayCPUMemory_Usage, and related monitoring tools) require the database's superuser role, so they are not available to a regularly-scoped database user. It requires TigerGraph version 4.2.0 or higher and Python 3.12 or higher, so older TigerGraph deployments are out of scope.

Capabilities

schema_introspection retrieves the complete vertex and edge type schema; query_view lists available queries and query_execution runs installed GSQL queries or raw GSQL with parameters. create_vertex_edge and upsert_vertex_edge create and update vertices/edges programmatically, including vector attributes, while alter_vertex enables vector-attribute support on an existing vertex type. udf_algorithm_listing surfaces installed user-defined functions and the GDS algorithm catalog. Query results can be exported to CSV or JSON.

How to install

git clone https://github.com/customdiscoveries/TigerGraph_MCP.git
cd TigerGraph-MCP
cp mcp_server/.env-example .env
cp mcp_server/mcp_chatbot/example-server_config.json server_config.json
installUVEnvironment.sh

Required environment variables are TG_HOST, TG_GRAPH, TG_USERNAME, TG_PASSWORD, and LLM_MODEL_FAMILY (ANTHROPIC or OPENAI); TG_SECRET and TG_TOKEN are auto-retrieved if not set, and the corresponding ANTHROPIC_API_KEY/OPENAI_API_KEY (plus model name and token limit variables) configure the chat-bot's LLM. Claude Desktop configuration points at the cloned main.py via its full path.

Who it's for

Developers and data teams working with TigerGraph who want natural-language access to schema inspection, query execution, and graph data mutation - including vector-attribute workflows for similarity search - plus database administrators who need superuser-gated visibility into service status, component versions, and resource usage across the TigerGraph cluster. Its built-in chat-bot mode, with commands like @listQueries to browse query result files, /tools to list available MCP tools, and /prompts to list available prompts, gives a lighter-weight interactive path for exploring a graph before wiring the same tools into a larger agent workflow, and query results can be exported to CSV or JSON for further analysis outside the chat session.

Source README

A community Python MCP server that provides TigerGraph database operations (queries, schemas, vertices, edges, user-defined functions) as structured tools for natural language interaction with TigerGraph databases.

Installation

From Source Code

git clone https://github.com/customdiscoveries/TigerGraph_MCP.git
cd TigerGraph-MCP
cp mcp_server/.env-example .env
cp mcp_server/mcp_chatbot/example-server_config.json server_config.json
installUVEnvironment.sh

Configuration

Claude Desktop

{
  "args": ["run", "/**Your full path goes here**/MCP-Repository/TigerGraph-MCP/main.py"]
}

Available Tools

Tool Description
schema_introspection Retrieve the complete graph schema (vertex and edge types)
query_view Get a list of available queries that can be executed
query_execution Execute installed GSQL queries or raw GSQL strings with parameters
create_vertex_edge Programmatically create vertices and edges
alter_vertex Modify a vertex to enable vector attribute support
upsert_vertex_edge Programmatically update vertex and edge attributes (including vector attributes)
udf_algorithm_listing Retrieve installed user-defined functions and GDS algorithm catalogs
displayService_Status Display database service status, showing which services are online (admin only)
displayDetailed_Service_Status Provide detailed service status with additional graph information, including vertices...
displayComponent_Version Display all component versions that make up the TigerGraph server (admin only)
displayCPUMemory_Usage Display CPU and memory usage by all TigerGraph server components (admin only)
displayDiskSpace_Usage Display disk space usage by various TigerGraph server components...

Features

  • Schema introspection - retrieve the complete graph schema (vertex and edge types)
  • Query viewing - get a list of available queries
  • Query execution - run installed GSQL queries or raw GSQL strings with parameters
  • Vertex and edge creation - programmatically create vertices and edges
  • Vertex modification - support for vector attributes
  • Vertex and edge updates - programmatically update vertex and edge attributes (including vector attributes)
  • User-defined functions and algorithms listing - retrieve installed user-defined functions and GDS algorithm catalogs
  • Administrative functions - monitor database services, component versions, CPU/memory/disk usage (requires 'superuser' role)
  • MCP chat-bot - interactive chat interface with natural language commands
  • CSV/JSON export - export query results to CSV or JSON formats

Environment Variables

Required

  • TG_HOST - TigerGraph host URL
  • TG_GRAPH - TigerGraph graph name
  • TG_USERNAME - TigerGraph database username
  • TG_PASSWORD - TigerGraph database password
  • LLM_MODEL_FAMILY - LLM model family - currently supports 'ANTHROPIC' or 'OPENAI'

Optional

  • TG_SECRET - TigerGraph secret (auto-retrieved if not specified)
  • TG_TOKEN - TigerGraph token (auto-retrieved if not specified)
  • ANTHROPIC_LLM_MODEL - Anthropic LLM model name
  • ANTHROPIC_API_KEY - Anthropic API key
  • ANTHROPIC_TOKENS - Anthropic token limit
  • OPENAI_LLM_MODEL - OpenAI LLM model name
  • OPENAI_TOKENS - OpenAI token limit
  • OPENAI_API_KEY - OpenAI API key

Usage Examples

@listQueries to see available query result files
@<query_file_name> list the contents of a query file
/tools to list available tools in the MCP server
/prompts to list available prompts in the MCP server
Show me the graph schema

Notes

Requires TigerGraph version >= 4.2.0 and Python >= 3.12. Administrative functions require the 'superuser' database role. Includes built-in chat-bot functionality and session management with automatic SECRET and TOKEN handling. Supports both Anthropic Claude and OpenAI LLM models for the chat-bot interface.

FAQ

Common questions

Discussion

Questions & comments · 0

Sign In Sign in to leave a comment.