FloydLabs
MCP Server Live

Connect Your LLM to Floyd Labs

73 production-ready AI skills organized across 3 MCP servers. Access them via the JSON-RPC MCP protocol, the REST API, or interactive Swagger docs.

3 MCP Servers • 67 Mapped Tools • 73 Total Skills

floyd-core

19 tools

Development operations, code analysis, build tools, git operations

Code AnalysisBuild AutomationGit OperationsTesting
typescript-semantic-analyzer, build-error-correlator, git-bisect, schema-migrator

ai-cognition

22 tools

AI reasoning, pattern recognition, context management

Pattern ExtractionSemantic AnalysisKnowledge SynthesisContext Management
concept-crystallization, pattern-synthesis, knowledge-graph-building, reasoning-chain-builder

ai-orchestration

26 tools

Multi-agent coordination, task management, resource allocation

Agent CoordinationTask ManagementResource AllocationConsensus Protocols
swarm-intelligence, workflow-orchestrator, conflict-resolver, emergent-behavior-detector

+ 10 Ghost Algorithms (bloom-sentinel, clone-lens, concept-lattice, consensus-voter, grammar-gate, merge-engine, patch-oracle, refactor-pathfinder, token-alchemist, viterbi-resolver) and 6 standalone analysis skills.

73
Total Skills
3
MCP Servers
10
Ghost Algorithms
$0
Cost

Two Ways to Access Floyd Labs

MCP Protocol (JSON-RPC 2.0)

Use the single-tool proxy pattern. Your LLM sees one tool ("floyd") that routes to all 73 skills. Context-efficient: 1 tool schema instead of 73.

POST https://floydslabs.com/api/mcp

REST API + Swagger Docs

Browse servers, list tools, execute skills, and check metrics via standard REST endpoints. Full interactive documentation available.

Open API Docs
1

Get Your API Key

Request access to get your unique Floyd Labs API key. Keys support rate limiting and usage tracking.

Request API Access
2

Configure Your MCP Client

Add Floyd Labs to your MCP client configuration.

Claude Desktop

Add to your Claude Desktop config file (claude_desktop_config.json):

{
  "mcpServers": {
    "floyd-labs": {
      "command": "npx",
      "args": ["-y", "@anthropics/mcp-proxy"],
      "env": {
        "MCP_PROXY_URL": "https://floydslabs.com/api/mcp",
        "MCP_PROXY_HEADERS": "Authorization: Bearer YOUR_API_KEY"
      }
    }
  }
}

Direct HTTP (Any Client)

For custom integrations, use the JSON-RPC 2.0 endpoint directly:

POST https://floydslabs.com/api/mcp
Content-Type: application/json
Authorization: Bearer YOUR_API_KEY

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/list",
  "params": {}
}
3

Start Using Tools

Once connected, your LLM can access all 73 skills through the proxy tool, or use the REST API:

// MCP: Execute a skill through the floyd proxy
{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "floyd",
    "arguments": {
      "action": "execute",
      "skill": "concept-crystallization",
      "args": { "action": "crystallize", "concept": "microservices" }
    }
  }
}

// REST: List all skills
GET /api/mcp/skills

// REST: Execute a skill directly
POST /api/mcp/skills/build-error-correlator/execute
Body: { "input": { ... } }

// REST: Browse server tools
GET /api/mcp/servers/floyd-core/tools

API Reference

Full interactive docs available at /api-docs

MCP Protocol (JSON-RPC)

initializeInitialize the MCP connection
tools/listList the floyd proxy tool
Auth
tools/callExecute a skill via the proxy
Auth
pingHealth check

REST API

POST /api/mcp/authAuthenticate (shared password → JWT)
GET /api/mcp/serversList MCP servers
GET /api/mcp/servers/{name}/toolsList tools per server
GET /api/mcp/skillsList all 73 skills
GET /api/mcp/skills/{name}Get skill details
POST /api/mcp/skills/{name}/executeExecute a skill
Auth
GET /api/mcp/healthHealth check
GET /api/mcp/metricsUsage metrics

The Floyd Labs Ecosystem

Ready to connect your AI to Floyd Labs?