MilvusSAFE
Model Context Protocol Servers for Milvus
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
The Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you're building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need.
This repository contains a MCP server that provides access to Milvus vector database functionality.
Prerequisites
Before using this MCP server, ensure you have:
- Python 3.10 or higher
- A running Milvus instance (local or remote)
- uv installed (recommended for running the server)
Usage
The recommended way to use this MCP server is to run it directly with uv without installation. This is how both Claude Desktop and Cursor are configured to use it in the examples below.
If you want to clone the repository:
git clone https://github.com/zilliztech/mcp-server-milvus.git cd mcp-server-milvus
Then you can run the server directly:
uv run src/mcp_server_milvus/server.py --milvus-uri http://localhost:19530
Alternatively you can change the .env file in the src/mcp_server_milvus/ directory to set the environment variables and run the server with the following command:
uv run src/mcp_server_milvus/server.py
Important: the .env file will have higher priority than the command line arguments.
Running Modes
The server supports two running modes: stdio (default) and SSE (Server-Sent Events).
Stdio Mode (Default)
- Description: Communicates with the client via standard input/output. This is the default mode if no mode is specified.
- Usage:
uv run src/mcp_server_milvus/server.py --milvus-uri http://localhost:19530
SSE Mode
- Description: Uses HTTP Server-Sent Events for communication. This mode allows
7b7f63eaa90fOBSERVED · 2026-10-06Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add mcp-server-milvus --env MILVUS_TOKEN=${MILVUS_TOKEN} -- uvx mcp-server-milvus{
"mcpServers": {
"mcp-server-milvus": {
"command": "uvx",
"args": [
"mcp-server-milvus"
],
"env": {
"MILVUS_TOKEN": "${MILVUS_TOKEN}"
}
}
}
}Exposed tools (14)
11 read · 2 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
milvus_create_collection | write | |
milvus_delete_entities | destructive | |
milvus_get_collection_info | read | |
milvus_hybrid_search | read | |
milvus_insert_data | write | |
milvus_list_collections | read | List all collections in the database. |
milvus_list_databases | read | List all databases in the Milvus instance. |
milvus_load_collection | read | |
milvus_query | read | |
milvus_release_collection | read | |
milvus_text_search | read | |
milvus_text_similarity_search | read | |
milvus_use_database | read | |
milvus_vector_search | read |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (2 observation(s))
- Network
- declared (1 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (6)
milvus_delete_entities
🔍 MCP Inspector is up and running at http://127.0.0.1:6274 🚀
You can then access the MCP Inspector at `http://127.0.0.1:6274` for testing.
"http://127.0.0.1:19530"
fastmcp, pymilvus, click, python-dotenv
Claude_mcp+1080.gif
Gates applied: no_behavioural_pass.
7b7f63eaa90ffull audit observations/trust-audit/mcp-server/zilliztech__milvus.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-06 | 7b7f63eaa90f | SAFE | B | 89 | first audit |
Questions
What is the Milvus MCP server?
Model Context Protocol Servers for Milvus
What tools does Milvus expose?
14 in total: 11 read-only, 2 that write, and 1 that can delete or overwrite (milvus_delete_entities). Every one is listed on this page with its risk.
Is Milvus safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Milvus need?
It reads MILVUS_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Milvus run?
It speaks sse and streamable-http, so it runs as a service you connect to over the network. It is published on PyPI as mcp-server-milvus.
How current is this page?
The grade is for one exact copy of the source (7b7f63eaa90f), read on 2026-10-06. The repository is watched and re-audited when it changes.