Try MidOS

Test MidOS tools instantly using the live cloud endpoint. No install, no API key, no signup.

Live Endpoint

MidOS is hosted on Smithery and available at:

https://midos-mcp--midosresearch.run.tools

You can call any free-tier tool via JSON-RPC. Every example below is a real request you can copy and run.

Example 1: Search Knowledge

Find knowledge chunks across 20+ technology stacks:

curl -s https://midos-mcp--midosresearch.run.tools \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "search_knowledge",
      "arguments": {
        "query": "React Server Components patterns",
        "max_results": 3
      }
    },
    "id": 1
  }'

What you get: Matching knowledge chunks with title, relevance score, stack, and content preview (truncated to 250 chars on free tier).

Example 2: List Skills by Stack

Browse available skills for any technology:

curl -s https://midos-mcp--midosresearch.run.tools \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "list_skills",
      "arguments": {
        "stack": "python"
      }
    },
    "id": 2
  }'

What you get: All Python-related skills with names, descriptions, and section counts. Try other stacks: typescript, go, rust, kubernetes, security.

Example 3: Agent Handshake

Tell MidOS about your agent to get personalized recommendations:

curl -s https://midos-mcp--midosresearch.run.tools \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "agent_handshake",
      "arguments": {
        "model": "claude-sonnet-4-5-20250514",
        "context_window": 200000,
        "client": "cursor",
        "languages": "python,typescript",
        "frameworks": "fastapi,nextjs",
        "project_goal": "Building a REST API with auth"
      }
    },
    "id": 3
  }'

What you get: Personalized tool catalog ranked by relevance, context window tips, and suggested skill packs for your exact stack.

Example 4: System Status

Check the knowledge base health and metrics:

curl -s https://midos-mcp--midosresearch.run.tools \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "hive_status",
      "arguments": {}
    },
    "id": 4
  }'

What you get: Live counts of chunks, skills, EUREKA entries, vector store status, and pipeline health.

From Terminal to MCP Client

Once you’ve seen the responses, connect MidOS to your AI agent for the full experience:

Install locally
Terminal window
```
pip install midos-mcp
```

Add to your MCP client (Claude Desktop, Cursor, VS Code, etc.)

{
  "mcpServers": {
    "midos": {
      "command": "midos-mcp",
      "args": ["serve"]
    }
  }
}

Start asking your agent

“Use search_knowledge to find patterns for JWT authentication”

Your agent will call MidOS tools automatically and get structured knowledge back.

See the full Quick Start for client-specific setup instructions.

Free Tier Tools

These 9 tools are available without an API key:

Tool	What it does
`search_knowledge`	Keyword search across 21,000+ knowledge chunks
`list_skills`	Browse 118 skills by technology stack
`hive_status`	Knowledge base health and metrics
`project_status`	System overview and pipeline status
`agent_handshake`	Personalized onboarding and tool recommendations
`where_was_i`	Resume a previous session
`context_health`	Monitor context window utilization
`get_progress`	Retrieve saved session progress
`agent_bootstrap`	Quick-start configuration for new agents

Want more? The Dev tier ($24/mo) unlocks all 34+ tools, full untruncated knowledge, EUREKA discoveries, semantic search, and 500 queries/day.