The Model Context Protocol (MCP) is an open standard that enables large language models (LLMs) to interact securely with external tools and data sources. MCP is like a universal plug, a USB-C port for artificial intelligence (AI). Instead of building fragile, custom integrations for every database, MCP provides a standardized way for AI models to connect directly to your database. This allows the model to understand schemas, execute queries, and retrieve real-time context without the need for data to leave the secure environment.
Why Use MCP for Agentic AI?
Beyond simple chatbots, agentic AI systems autonomously reason, plan, and execute tasks. These agents require secure, direct access to a data estate rather than a mere training dataset to remain effective.
Autonomous Exploration
MCP allows agents to explore your database schemas to find the right data to answer a complex prompt, reducing the need for hard-coded prompts.
Grounded Truth
By querying Actian Ingres or HCL Informix® in real time, agents base their answers on the latest business facts. This directly reduces AI hallucinations.
Tool-Use Capabilities
MCP transforms database functions (such as a complex join in the Analytics Engine or a search in NoSQL) into tools the agent can intelligently select and use when appropriate.
Contextual Memory
Agents can use edge databases, like Actian Zen, to store and retrieve long-term states across different user sessions.
The Actian MCP Hub: Specialized Connectors
While MCP is a universal protocol, performance must be native: we provide a centralized Docker Hub repository with optimized images for our entire database portfolio.
| Database | MCP Image Name | Suited For |
|---|---|---|
| Actian Ingres | actian/ingres-mcp-server | Mission-critical relational data and enterprise logic |
| HCL Informix® | actian/informix-mcp-server | Time-series, Internet of Things (IoT), and high-availability spatial data |
| Actian Zen | actian/zen-mcp-server | Edge-based AI and zero-admin mobile applications |
| Actian NoSQL | actian/nsql-mcp-server | High-fidelity context from complex object structures |
| Actian Analytics Engine | actian/analytics-engine-mcp-server | Large-scale retrieval-augmented generation (RAG) and high-performance vector analysis |
Enterprise Architecture
The Actian MCP Server acts as a stateless security proxy between the AI agent and the database
1. Agent Request
The AI agent (for example, Claude or GPT-4o) requests a specific tool or data point.
2. Native Translation
The MCP Server, running securely in your Docker environment, translates that request into native database syntax.
3. Secure Execution
The database executes the request and returns only the necessary context.
4. Agentic Action
The agent uses that context to complete its task or plan its next move.
Get Started in Minutes
All MCP server images are available at hub.docker.com/u/actian. You can deploy a specific server quickly using standard container orchestration.
# Example: Launch the MCP Server for HCL Informix®
docker pull actian/informix-mcp-server:latest
# Run with your secure environment variables
docker run -e INFORMIXSERVER=myserver \
-e DB_NAME=finance \
actian/informix-mcp-serverSecurity and Governance
Identity-Aware Access
The server's modern authentication protocols ensure agents access only explicitly authorized data.
Set Up Authentication →Enforced Read-only Mode
Configuration toggles guarantee that AI agents can read data but can never modify or delete it.
Learn about Read-only Mode →Comprehensive Audit Trails
The system logs every tool call and query generated by the AI for full compliance and monitoring.
Explore the Architecture →