Oso Cloud MCP Server
Understand, develop, and debug authorization policies in Oso Cloud.
Coding agents are everywhere. See what yours are doing →
Local Development
Oso MCP Server
Use the MCP Server to:
- develop a better understanding of your policy
- use natural language to ask authorization-related questions
- debug why an authorization decision does not match what you expect
This guide assumes you have already installed and set up the CLI.
Connect
Claude Desktop (DXT)
Don’t have the CLI installed or don’t want to manually add the configuration? No worries - download the following DXT file and double-click to automatically install the Oso MCP Server into Claude Desktop or other AI clients that accept DXTs.
curl -O https://d2nl1ljmg8n1q8.cloudfront.net/latest/oso-cloud-mcp.dxt
Verify the SHA256 checksum: MacOS:
curl -sL https://d2nl1ljmg8n1q8.cloudfront.net/latest/oso-cloud-mcp.dxt.sha256 | shasum -a 256 -c
Linux:
curl -sL https://d2nl1ljmg8n1q8.cloudfront.net/latest/oso-cloud-mcp.dxt.sha256 | sha256sum -c
You may also specify the DXT version:
curl -O https://d2nl1ljmg8n1q8.cloudfront.net/0.33.2/oso-cloud-mcp.dxt
Note: There are no DXT versions <0.33.0 The following installation methods assume you have already installed and set up the CLI (version >=0.33.0).
Cursor
Manual
In your LLM client of choice, add the following to your MCP config file to run it against a local dev server:
{
"mcpServers": {
"oso": {
"command": "oso-cloud",
"args": ["experimental", "mcp"],
"env": {
"OSO_URL": "http://localhost:8080",
"OSO_AUTH": "e_0123456789_12345_osotesttoken01xiIn"
}
}
}
}
- You may optionally omit the
OSO_URLenvironment variable or set it tohttps://cloud.osohq.comif you want to run it against a live production server. - The provided
OSO_AUTHtoken is for the local dev server. You can obtain your live serverOSO_AUTHtoken from the Oso Cloud UI.
Where can I find my MCP config file?
Here are guides for some common clients:
- Claude Desktop
- Cursor
- VS Code Copilot
- Kiro
Usage
This server is primarily intended for use with dev servers to aid in development. Please use extra caution when using against a live environment.
Once you have your MCP server up and running, you can ask your LLM any authorization related questions and watch it use the tools available. Currently, we expose read tools to:
- Get your policy
- Get all facts
- Run an
authorizequery - Query your facts with pattern matching
- Run policy tests We also expose the following write tools, restricted for use only with local dev servers:
- Update your policy
- Add facts
- Delete facts
- Clear all data Try sending any of the following messages:
- Draw a mermaid diagram of my authorization policy
- What permissions does
<actor>have on<resource>?- e.g. “What permissions does Alice have on Project XYZ?”
- Why doesn’t
<actor>have permission to<action><resource>?
Feedback
We are actively iterating on developer experience and would appreciate all feedback on the Oso MCP Server and the broader development experience with Oso Cloud. Please do not hesitate to reach out on Slack!
Was this page helpful?
YesNo
Oso Migrate (Beta)
Overview of Polar Language and Syntax
Verwandte Server
Scout Monitoring MCP
SponsorPut performance and error data directly in the hands of your AI assistant.
Alpha Vantage MCP Server
SponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Remote MCP Server on Cloudflare
A remote MCP server deployable on Cloudflare Workers with OAuth login support, designed for both local development and cloud deployment.
MCP Playground
A playground for MCP implementations featuring multiple microservices, including news and weather examples.
FastMCP-Scala
A Scala 3 library for building Model Context Protocol (MCP) servers.
Qwen-Agent
A framework for developing LLM applications with capabilities like tool usage, planning, and memory, based on the Qwen model.
Agent Evals by Galileo
Bring agent evaluations, observability, and synthetic test set generation directly into your IDE for free with Galileo's new MCP server
Svelte Documentation
Remote server (SSE/Streamable) for the latest Svelte and SvelteKit documentation
Azure DevOps
Integrate with Azure DevOps services to manage work items, repositories, and pipelines.
Pathmode
Build structured intent specs through Socratic AI conversation. Describe a problem or paste a support ticket — Claude challenges vague thinking, asks pointed questions, and builds a structured spec. Exports as intent.md, .cursorrules, or CLAUDE.md.
MCP Agentic Development Platform
A comprehensive MCP development environment with interactive visualizations, multiple client interfaces, and advanced agentic capabilities.
tachibot-mcp
Stop AI Hallucinations Before They Start Run models from OpenAI, Google, Anthropic, xAI, Perplexity, and OpenRouter in parallel. They check each other's work, debate solutions, and catch errors before you see them.