Agent Fusion
Agent Fusion is a local RAG semantic search engine that gives AI agents instant access to your code, documentation (Markdown, Word, PDF). Query your codebase from code agents without hallucinations. Runs 100% locally, includes a lightweight embedding
What is Agent Fusion?
Agent Fusion is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to agent fusion is a local rag semantic search engine that gives ai agents instant access to your code, documentation (markdown, word, pdf). query your codebase from code agents without hallucinations. r...
Agent Fusion is a local RAG semantic search engine that gives AI agents instant access to your code, documentation (Markdown, Word, PDF). Query your codebase from code agents without hallucinations. Runs 100% locally, includes a lightweight embedding
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Agent Fusion is a local RAG semantic search engine that give
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx agent-fusionConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Agent Fusion
Agent Fusion is a local RAG (Retrieval-Augmented Generation) semantic search engine packaged as an MCP server, giving AI coding agents instant access to your codebase and documentation without sending any data to the cloud. It indexes Markdown, Word, PDF, and source code files into a local DuckDB database and exposes a single powerful query_context tool that runs semantic search, symbol matching, and full-text keyword search in parallel. The server runs as a self-contained JAR file that requires Java 21 and no external API keys.
Prerequisites
- Java 21 or later installed (download from adoptium.net)
- At least one directory of code, Markdown, PDF, or DOCX files to index
- An MCP-compatible client such as Claude Desktop, Claude Code, or Cursor
- The Agent Fusion release ZIP downloaded from the GitHub releases page
Download the Agent Fusion release
Go to https://github.com/krokozyab/Agent-Fusion/releases and download the latest release.zip. Extract it to a permanent location — you will find orchestrator-0.1.0-all.jar, a fusionagent.toml config file, and startup scripts.
Configure watch paths in fusionagent.toml
Edit fusionagent.toml and set watch_paths to the directories you want Agent Fusion to index. The file watcher automatically re-indexes changes without manual refresh.
watch_paths = [
"/Users/yourname/projects/my-app",
"/Users/yourname/docs"
]Start the Agent Fusion server
Run the startup script for your platform. The server starts an HTTP MCP endpoint on port 3000 and opens an admin dashboard in your browser.
# macOS / Linux
./start.sh
# Verify it is running
curl http://127.0.0.1:3000/mcpConfigure your MCP client to connect
Add the Agent Fusion server to your claude_desktop_config.json using the proxy script included in the release, or connect Claude Code directly via HTTP transport.
# Claude Code (HTTP transport)
claude mcp add --transport http orchestrator http://127.0.0.1:3000/mcp
# Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"agent-fusion": {
"command": "/path/to/orchestrator-mcp-proxy.sh",
"env": {
"MCP_ENDPOINT": "http://127.0.0.1:3000/mcp/json-rpc"
}
}
}
}Wait for initial indexing to complete
On first start Agent Fusion indexes all files in the configured watch_paths. Check the admin dashboard in your browser to track indexing progress before making queries.
Agent Fusion Examples
Client configuration
Claude Desktop config connecting to the locally running Agent Fusion MCP endpoint via the proxy script.
{
"mcpServers": {
"agent-fusion": {
"command": "/path/to/orchestrator-mcp-proxy.sh",
"env": {
"MCP_ENDPOINT": "http://127.0.0.1:3000/mcp/json-rpc"
}
}
}
}Prompts to try
Example queries to send to your AI assistant once Agent Fusion is indexing your codebase.
- "Use query_context to find all authentication logic in the codebase"
- "Search my indexed docs for anything about database connection pooling"
- "Find the error handling patterns used across the project"
- "Where is JWT validation implemented in my code?"
- "Search for all places where the UserService class is referenced"Troubleshooting Agent Fusion
Server fails to start with Java version error
Agent Fusion requires Java 21 or higher. Run 'java --version' to check. Download the latest LTS release from adoptium.net and ensure it is on your PATH before running start.sh.
query_context returns no results for files that exist
Check the admin dashboard to confirm the relevant directories are listed in watch_paths in fusionagent.toml and that initial indexing has completed. Also verify the file types are among the supported formats (.kt, .py, .ts, .java, .pdf, .docx, .md).
MCP client cannot connect to the server endpoint
Confirm the server is running by visiting http://127.0.0.1:3000 in a browser. Ensure the MCP_ENDPOINT in your config points to http://127.0.0.1:3000/mcp/json-rpc and that no firewall is blocking port 3000.
Frequently Asked Questions about Agent Fusion
What is Agent Fusion?
Agent Fusion is a Model Context Protocol (MCP) server that agent fusion is a local rag semantic search engine that gives ai agents instant access to your code, documentation (markdown, word, pdf). query your codebase from code agents without hallucinations. runs 100% locally, includes a lightweight embedding It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Agent Fusion?
Follow the installation instructions on the Agent Fusion GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Agent Fusion?
Agent Fusion works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Agent Fusion free to use?
Yes, Agent Fusion is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Agent Fusion Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to Agent Fusion? Here are other popular knowledge & memory servers you can use with Claude, Cursor, and VS Code.
MemPalace
★ 52.6kA local AI memory system that stores all conversations verbatim and organizes them into navigable structures. It provides 19 MCP tools for AI assistants to search and retrieve past decisions, debugging sessions, and architecture debates automatically
Kratos
★ 25.7k🏛️ Memory System for AI Coding Tools - Never explain your codebase again. MCP server with perfect project isolation, 95.8% context accuracy, and the Four Pillars Framework.
Context Mode
★ 15.4kAn MCP server that preserves LLM context by intercepting large data outputs and returning only concise summaries or relevant sections. It enables efficient sandboxed code execution, file processing, and documentation indexing across multiple programm
Memu
★ 13.7kMemory for 24/7 proactive agents like OpenClaw.
MemOS
★ 9.3kMemOS (Memory Operating System) is a memory management operating system designed for AI applications. Its goal is: to enable your AI system to have long-term memory like a human, not only remembering what users have said but also actively invoking, u
Everos
★ 5.4kBuild, evaluate, and integrate long-term memory for self-evolving agents.
Browse More Knowledge & Memory MCP Servers
Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Agent Fusion in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Agent Fusion?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.