research

What is research?

research is a Quarkus-based application that uses the Model Context Protocol (MCP) and LangChain4j to implement a simple agentic assistant. It runs on the Quarkus Java framework and is intended for developers building multi‑MCP‑server agents.

How to use research?

Install Node.js, npm, and a container runtime (Podman/Docker) for optional telemetry. Create a playground directory for the filesystem MCP server. Obtain API keys for Brave Search, OpenAI, Google Maps, and Slack. Set them in a .env file (or application.properties for testing). Run ./mvnw compile quarkus:dev to start in dev mode; access the chat UI at http://localhost:8080 or through the Quarkus Dev UI.

Key features of research

• Orchestrates multiple MCP servers (Brave Search, Google Maps, Slack, filesystem)
• Built with Quarkus and LangChain4j for Java
• Provides both a web chat UI and a Dev UI chat interface
• Supports packaging as a JAR or native executable
• Includes built‑in telemetry via Grafana (requires a container)

Use cases of research

• Find a restaurant that matches dietary restrictions and preferences for a group.
• Send restaurant recommendations to a Slack channel.
• Create an ICS calendar file for an event in a local directory.
• Explain the reasoning steps and tool invocations used to arrive at a recommendation.

FAQ from research

What is the research application?

It is a Quarkus + LangChain4j app that uses the Model Context Protocol to combine multiple MCP servers into a single agentic assistant. It runs in Java and can be packaged as a JAR or native binary.

Which MCP servers does research support?

The example uses Brave Search, Google Maps, Slack, and a filesystem server. Additional MCP servers can be configured via application.properties.

What platforms or models does research run on?

It runs on any platform that supports Quarkus (Java 17+). The LLM backend is OpenAI (via LangChain4j), configurable through an API key.

How do I obtain the required API keys?

Links in the README point to the official pages for each service: Brave Search API, OpenAI API, Google Maps API, and Slack Bot User OAuth Token. Keys are set in a .env file or application.properties.

Is research free or open source?

The README does not specify licensing, but the project is hosted on GitHub and uses open‑source frameworks (Quarkus, LangChain4j). Usage of external services (Brave, OpenAI, etc.) requires their own API keys and may incur costs.