Agent-Fusion

Overview

Agent Fusion is a local RAG semantic search engine that gives AI agents instant access to your code, documentation, and PDFs through the MCP protocol. It consists of two main components: a Context Engine for intelligent indexing and search, and an optional Task Manager for coordinating multiple AI agents. The system runs entirely locally with no external dependencies, using a lightweight embedding model that can be replaced with more powerful options. It provides semantic, symbol, and full-text search capabilities to help agents find relevant code and documentation without hallucinations.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• query_context

  Searches indexed code, documents, and PDFs using semantic, symbol, and full-text search methods

  Try:Use query_context to find authentication logic in the codebaseTry:Search for database connection pooling patterns using query_contextTry:Find error handling implementations with query_context

Comparable tools

Installation

{
  "mcpServers": {
    "agent-fusion": {
      "command": "java",
      "args": ["-jar", "/path/to/agent-fusion.jar"],
      "env": {}
    }
  }
}

FAQ

What file types does Agent Fusion support?

Agent Fusion supports code files (.kt, .py, .ts, .java), documents (.pdf, .docx, .md), and any custom file types you configure in the configuration file.

Can I use Agent Fusion without the Task Manager?

Yes, the Context Engine can be used independently for semantic search without the Task Manager component.

Compare Agent-Fusion with