MCP servers · least privilege

Read-only MCP servers

1,620 MCP servers declare tools that only read: none of their tools is classed write or destructive. That makes them the safest place to start giving an agent access. Read-only describes the tools, not the code, so the audit verdict still decides whether to connect one.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “Read-only MCP servers”. Retrieved 29 Oct 2026. https://modelsatlas.com/mcp/read-only/page/28

Verdicts · 1,620 servers

✓SAFE913!CAUTION261✕BLOCK446

Transports

A server can list more than one transport.

1,620 servers

649–672 of 1,620 · ranked by trust score, then stars

Obsidian Omnisearch

anpigon/mcp-server-obsidian-omnisearch

✓SAFE B

An MCP server that enables searches within Obsidian vaults using the Omnisearch plugin API. Returns absolute paths to matching notes for seamless integration. 🚀

★ 302 read

Outsource

gwbischof/outsource-mcp

✓SAFE B

Give your AI assistant its own AI assistants.

★ 302 readstdio

PM Copilot

dkships/pm-copilot

✓SAFE B

MCP server that triangulates customer support tickets and feature requests to help PMs decide what to build next

★ 305 readstdio

Pulse

orliesaurus/pulsemcp-server

✓SAFE B

An MCP server that helps you find MCP servers that are listed on PulseMCP.com

★ 302 readstdio

Skills

skills-mcp/skills-mcp

✓SAFE B

Bring Claude's Skills pattern to any MCP-compatible agent

★ 302 readstdio

3GPP

edhijlu/3gpp-mcp-server

✓SAFE B
★ 2918 readstdio

Android Code Search

steveday763/cs_android_mcp

✓SAFE B
★ 2910 readstdio

Claude Auto-Approve

pynesys/claude_autoapprove_mcp

✓SAFE B

An MCP server to inject auto-approve MCP functionality into Claude Desktop

★ 292 read

Cowork Semantic Search

zhubit/cowork-semantic-search

✓SAFE B
★ 291 read

Demo Implementation

mschwarzmueller/demo-mcp-server-client-implementation

✓SAFE B

A demo implementation of a MCP server (consuming a dummy API) and basic client.

★ 291 readstdio

Excel Reader

archimedescrypto/excel-reader-mcp

✓SAFE B

A Model Context Protocol (MCP) server for reading Excel files with automatic chunking and pagination support. Built with SheetJS and TypeScript.

★ 2914 readstdio

Forensic Toolkit

axdithyaxo/mcp-forensic-toolkit

✓SAFE B

Secure AI-enabled digital forensics toolkit built with Model Context Protocol

★ 295 read

iMessage

wyattjoh/imessage-mcp

✓SAFE B

A Model Context Protocol server for reading iMessage data from macOS.

★ 296 readstdio

Jina AI Reader

spences10/mcp-jinaai-reader

✓SAFE B

🔍 Model Context Protocol (MCP) tool for parsing websites using the Jina.ai Reader

★ 291 readstdio

Linkup

linkupplatform/linkup-mcp-server

✓SAFE B

Linkup MCP server for web search and fetching web content

★ 294 readstdio

Magento 2 LSP

mage-os-lab/magento2-lsp

✓SAFE B

Language Server (LSP) and MCP server for Magento 2 - provides go-to-definition, references, and rename across PHP, XML, and layout files.

★ 2911 readstdio

Moneybird

vanderheijden86/moneybird-mcp-server

✓SAFE B

Moneybird MCP Server - AI-powered bookkeeping

★ 292 readstdio

Music Analysis

hugohow/mcp-music-analysis

✓SAFE B

Integrate librosa, whisper with LLMs to analyze music audio.

★ 298 read

RedNote

milescool/rednote-mcp

✓SAFE B

RedNote MCP - Xiaohongshu Content Search Tool

★ 291 readstdio

Textin

intsig-textin/textin-mcp

✓SAFE B
★ 294 readstdio

Agrobr

bruno-portfolio/agrobr-mcp

✓SAFE B

MCP server for Brazilian agricultural data — connect LLMs to 10 public data sources via agrobr

★ 2810 readstdio

Baidu Search

evilran/baidu-mcp-server

✓SAFE B

Baidu Search MCP Server I A Model Context Protocol (MCP) server that provides web search capabilities through Baidu, with additional features for content fetching and parsing.

★ 281 read

Bonnard

bonnard-data/bonnard

✓SAFE B

Archived. Earlier Bonnard product. Current: bonnard.dev / @bonnard/mcp-charts

★ 286 readstreamable-http

Jina AI

joebuildsstuff/mcp-jina-ai

✓SAFE B
★ 283 readstdio

Other risk hubs

By job

Questions

What does read-only mean here?
A server is listed here when its source declares at least one tool and every tool is classed read: no write tools and no destructive tools.
Is a read-only server always safe?
No. 913 are SAFE, 261 CAUTION and 446 BLOCK on this list. A tool list says what the tools do, not whether the surrounding code is clean, so a read-only server can still be BLOCK.
Why are some servers on no risk list?
615 servers declare no tools we can read, because they build the list at runtime. They cannot be called read-only, so they appear on none of the risk hubs.

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.