Klavis AI

{
  "service_type": "platform",
  "base_url": "https://klavis.ai",
  "auth_method": "oauth",
  "auth_config": {
    "type": "oauth2",
    "scopes": ["openid", "profile", "email"],
    "authorization_url": "https://klavis.ai/oauth/authorize",
    "token_url": "https://klavis.ai/oauth/token"
  },
  "endpoints": [
    {
      "path": "/api/v1/mcp/servers",
      "method": "GET",
      "description": "List hosted MCP servers available to the user"
    },
    {
      "path": "/api/v1/mcp/servers/{serverId}",
      "method": "POST",
      "description": "Launch or provision an MCP server instance"
    },
    {
      "path": "/api/v1/mcp/connect",
      "method": "POST",
      "description": "Establish a session to a remote MCP server"
    },
    {
      "path": "/api/v1/mcp/users",
      "method": "GET",
      "description": "Retrieve user context and multi-tenant information"
    },
    {
      "path": "/api/v1/integrations/jira",
      "method": "POST",
      "description": "Trigger Jira integration setup or data exchange"
    }
  ],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "note": "Pricing not public; likely freemium with a free tier and paid plans for hosted MCP servers and scale. Specific pricing model not documented in provided content."
    }
  },
  "rate_limits": {
    "note": "Not disclosed publicly in available documentation."
  },
  "capabilities": [
    "Hosted MCP server provisioning",
    "OAuth and multi-tenancy authentication",
    "White-labeled OAuth integration",
    "Remote MCP connection API for backend integration",
    "Pre-built web, Slack, and Discord clients",
    "SDKs for Python and TypeScript",
    "Supported integrations (e.g., Jira, Slack, Discord)",
    "Open source under MIT license"
  ],
  "raw_analysis": "Klavis AI is an open-source platform providing hosted MCP (Model Context Protocol) servers for AI applications. It aims to simplify the integration of AI agents with external tools like Jira, Slack, and Discord. The platform is early-stage (founded 2025, Y Combinator backed) and offers enterprise-grade features such as OAuth and multi-tenancy, eliminating the need for heavy client-side development. Maturity is low due to recent founding, but the open-source nature (MIT license) and YC backing provide some credibility. The platform exposes a REST API for managing and connecting to MCP servers, with SDKs in Python and TypeScript, and pre-built clients for web, Slack, and Discord. Key capabilities include hosted MCP provisioning, secure authentication, and support for popular integrations. Pricing is likely freemium but not publicly detailed. The service is positioned for developers building AI agents that need secure, scalable tool integrations without managing backend infrastructure."
}

3/3 tests passed

{
  "overall": 72,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 6.5,
    "response_parseability": 7.5,
    "error_clarity": 6.0,
    "doc_quality": 7.0,
    "auth_simplicity": 8.5,
    "latency": 8.0,
    "consistency": 7.5
  },
  "pricing_normalized": {
    "type": "freemium",
    "note": "Pricing not public; likely freemium with a free tier and paid plans for hosted MCP servers and scale."
  },
  "issues": [
    "Pricing is not publicly documented, making it hard to estimate cost for users.",
    "The robots.txt disallows /mcp-server-connections/*--*--* sections, which might hide some technical docs from crawlers.",
    "Limited public documentation beyond the llms.txt summary; no detailed pricing or feature comparisons found.",
    "No explicit error handling or limitation descriptions in the llms.txt content, reducing error clarity.",
    "No status page or uptime history visible, impacting consistency assessment."
  ],
  "recommendations": [
    "Publish clear pricing tiers and a free-tier limits page to aid agent recommendations.",
    "Provide structured API docs (OpenAPI) and example requests/responses for better parseability.",
    "Add a public status page and uptime history to boost user trust and consistency scoring.",
    "Document common errors and limitations in a FAQ or error code reference section.",
    "Include a 'getting started' guide with auth flow steps (OAuth setup) to improve first-try success."
  ]
}