OpenL MCP Server
OfficialThe OpenL MCP Server connects AI assistants to OpenL Studio β a Business Rules Management System β enabling them to manage repositories, projects, rules/tables, tests, traces, and deployments programmatically.
Repository & Project Management
List repositories, branches, features, and project revision history
List, open, save, close, and create projects; branch from specific revisions
Track and restore local (unsaved) project changes
File & Table (Rules) Operations
List, get, update, append, create, and delete tables (Decision Tables, Spreadsheets, Datatypes, Test tables, etc.)
Read, write, search, copy, move, and delete project files
Apply raw-source edits to tables
Testing
Start project-level or table-specific test runs
Retrieve results in summary, full, or table-specific formats
Tracing & Debugging
Start, cancel, and export traces
Navigate trace trees; inspect nodes, parameters, context, and errors
Load lazy-loaded parameter values
Deployment
List deploy repositories and active deployments
Deploy new projects or redeploy/rollback existing deployments
AI Prompts
14 expert guidance templates for complex workflows (e.g., creating rules, running tests, deploying)
Core Features: Type-safe validation (Zod), PAT-based authentication, pagination for list operations, and multiple response formats (JSON, Markdown variants).
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@OpenL MCP Serverlist my repositories"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
OpenL MCP Server
Model Context Protocol server for OpenL Studio Business Rules Management System.
Built with MCP SDK v1.29+ featuring type-safe validation (Zod) and comprehensive OpenL Studio integration.
Quick Links
π Quick Start - Connect Claude Code, Claude Desktop, Cursor, or VS Code in ~5 minutes
βοΈ Advanced Guide - Settings, authentication, Docker, and CLI mode
π₯οΈ CLI Guide - Use the same binary as a shell tool (no MCP client needed)
π Usage Examples - Learn how to use MCP tools
π Troubleshooting - Common issues and solutions
π¨βπ» Contributing - Development guide
Related MCP server: Operaton MCP
npm Distribution
The MCP server is published as an npm package: openl-mcp β stdio transport via npx for Claude Code / Claude Desktop / Cursor / VS Code. No Node.js? Run it on the official Docker image instead β see the Advanced Guide.
The same binary also doubles as a CLI for direct API calls without an MCP client (npx -y openl-mcp <tool> '<json-args>') β see the CLI Guide for the full reference.
For npm package details see README.npm.md.
Documentation Structure
Getting Started
Quick Start - Connect your AI client (Claude Code, Claude Desktop, Cursor, VS Code)
Guides
Usage Examples - Practical examples of using MCP tools
Advanced Guide - Settings, authentication, Docker, and CLI mode
CLI Guide - Use the same binary as a shell tool (no MCP client needed)
Troubleshooting Guide - Common issues, debugging, and solutions
Development
Contributing Guide - How to contribute to the project
Architecture - System architecture and design
Testing Guide - Testing strategy and how to run tests
Code Standards - Best practices and coding standards
Tool Review - Technical review of MCP tools vs OpenL API
OpenL Studio Concepts
OpenL Studio uses dual versioning: Git-based commits (temporal) and dimension properties (business context). Supports multiple table types: Decision Tables (Rules, SimpleRules, SmartRules, Lookups), Spreadsheet Tables, and others (Method, Datatype, Test, etc.).
See prompts/create_rule.md for detailed table type guidance.
Tools
The MCP server provides 58 tools for managing OpenL Studio repositories, projects, rules, tables, tests, traces, and deployments. All tools are prefixed with openl_ and versioned (v1.0.0+).
Categories:
Guidance - Agent onboarding (
openl_get_started), per-project AGENTS.md context (openl_get_project_agent_context), and the official OpenL reference documentation embedded at build time (openl_list_guides/openl_get_guides)Repository Management - List repositories, branches, features, and revisions
Project Management - List, open, save, close, create, and branch projects; track local changes
Files - Read, write, search, copy, move, and delete project files
Rules & Tables - List, get, update, append, create, and delete tables; apply single raw-source edits (insert/delete/update/merge a row, column, or cell)
Tests - Start tests and retrieve results (full, summary, or by table)
Tracing - Interactive rule debugger: start a session, step into/over/out, set breakpoints, run to a stop, inspect live frame variables and decision-table outcomes, profile a whole run
Deployment - List deploy repositories and deployments; deploy and redeploy projects
See Usage Examples for detailed tool usage.
Prompts
14 expert guidance templates for complex OpenL Studio workflows. Prompts provide contextual assistance, best practices, and step-by-step instructions directly in Claude Desktop or MCP Inspector.
Available prompts: local_projects, create_rule, create_rule_decision_tables, create_rule_spreadsheet, create_test, update_test, run_test, append_table, datatype_vocabulary, dimension_properties, deploy_project, validate_after_edit, project_agents_md, project_history.
Usage: Request prompts in Claude Desktop (e.g., "Use the create_rule prompt") or access via MCP Inspector. See prompts/create_rule.md for detailed content.
Configuration
Base URL
Pass the OpenL Studio base URL as a positional argument (preferred), or set OPENL_BASE_URL:
# Positional argument (preferred) β starts the stdio MCP server
openl-mcp http://localhost:8080
npx -y openl-mcp http://localhost:8080
# β¦or via the environment variable
OPENL_BASE_URL=http://localhost:8080 openl-mcpThe positional URL takes precedence over OPENL_BASE_URL if both are set.
Environment Variables
# Base URL (or pass it as the positional argument above)
OPENL_BASE_URL=<your-base-url>
# Auth is optional (single-user mode accepts unauthenticated requests).
# Personal Access Token
OPENL_PERSONAL_ACCESS_TOKEN=<your-token>
# Optional
OPENL_TIMEOUT=60000See the Advanced Guide for authentication and all settings.
Client configuration
See the Quick Start for client-specific configuration (Claude Code, Claude Desktop, Cursor, VS Code).
Key Features
Type-Safe: Zod schemas with strict validation and TypeScript inference
Personal Access Token Auth: PAT-based authentication (or none, for single-user mode)
4 Response Formats: json, markdown, markdown_concise, markdown_detailed
Pagination Support: Metadata for all list operations
AI Prompts: 14 expert guidance templates
Comprehensive Tests: Full test suite covering core functionality
Development
npm run build # Build TypeScript
npm test # Run all tests
npm run lint # Check code quality
npm run watch # Dev mode with auto-rebuildSee Contributing Guide for development guidelines and Testing Guide for test suites.
Project Structure
openl-mcp/
βββ src/ # Source code (TypeScript)
βββ tests/ # Jest test suites
βββ prompts/ # AI assistant guidance (OpenL-specific)
βββ dist/ # Compiled output
βββ docs/ # Documentation
β βββ guides/ # User guides: setup, usage, auth, troubleshooting, Docker
β βββ development/ # Developer documentation
βββ README.md # This fileMore Documentation
Documentation Index - Full navigation map
Resources
License
LGPL-3.0 - GNU Lesser General Public License v3.0 (follows OpenL Studio project license).
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
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/openl-tablets/openl-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server