Devrag: MCP server, setup and security

What Devrag does

According to its repository: “Local RAG for semantic vector search over Markdown documents with sqlite-vec and multilingual-e5-small embeddings, filterable by directory and filename.”

The Model Context Protocol (MCP) is a standard plug between an AI agent and an outside service: once the server is connected, its tools show up in the agent, which decides when to call them. Devrag is listed under Memory and knowledge, a category of 250 servers in the directory, where it ranks #96 by stars.

Topics declared by the repository: claude-code, developer-tools, documentation, embeddings, golang, local-ai.

Installing and connecting Devrag

No package or remote server is declared in the official MCP registry: you install it from source, following the repository README.

It works with any MCP-compatible client (Claude Desktop, Claude Code, Cursor, VS Code, Codex and most recent agents). You add one entry to the client's MCP configuration: server name, start command or URL, environment variables. Project language: Go. Listed systems: macOS, Windows, Linux.

Security: what to check

Connecting an MCP server means installing software that talks straight to your agent: it runs with your rights and the model trusts its tool descriptions. The risk therefore depends on what it can touch and how far its code can be trusted.

Points found for Devrag:

• No licence: without a licence the code is not legally reusable, even if it is public. Avoid it in a company setting without the author's agreement.

Good practice: grant only the rights needed (least privilege), require human approval for irreversible actions, and run the AI agent security check.

Project activity and maintenance

Created on October 24, 2025, the tomohiro-owada/devrag repository has 64 stars and 15 forks, with 0 open issues. It has more stars than 57% of the servers in the directory. Last change 6 months ago. Version published in the registry: 1.2.0 (December 15, 2025).

Stars measure attention, not quality: a recent, heavily promoted project can collect them fast, while a useful but niche server will have few.