Klavis AI

Klavis AI is an open-source MCP integration platform that lets AI agents reliably use external tools at production scale. The product line is organized around Strata (intelligent connectors that compress and route tool context), 100+ prebuilt MCP integrations with OAuth, and an MCP Sandbox for live agent training and evaluation. Target customers are AI agent companies, RL teams, and enterprises that need long-horizon multi-app environments with seeded state, resets, and verifiable outcomes, plus SOC 2 Type II and GDPR posture. SDKs are available for Python and TypeScript/JavaScript, integrations cover Claude, OpenAI, Gemini, Cohere, Mistral, LangChain/LangGraph, LlamaIndex, CrewAI, Mastra, Agno, Fireworks, Together, and Google ADK, and the project is Apache-2.0 on GitHub.

1 APIs 0 Features

MCPModel Context ProtocolMCP ServersMCP HostingConnectorsOAuthSandboxesAgent TrainingReinforcement LearningWhite LabelOpen SourceStrata

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Klavis AI MCP Platform API
  version: v1
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: MCP Servers
    type: folder
  items:
  - info:
      name: Create MCP Server Instance
      type: http
    http:
      method: POST
      url: https://api.klavis.ai/api/mcp-server/create
      body:
        type: json
        data: '{}'
    docs: Create a URL for a specified MCP server. Validates the request with an API key and user details, returning an instance
      identifier and connection URL.
- info:
    name: Tools
    type: folder
  items:
  - info:
      name: List MCP Server Tools
      type: http
    http:
      method: GET
      url: https://api.klavis.ai/api/mcp-server/:server_id/tools
      params:
      - name: server_id
        value: ''
        type: path
        description: Identifier of the MCP server instance
      - name: format
        value: ''
        type: query
        description: Tool schema format to return
    docs: List all tools available for a specific remote MCP server in various AI model formats.
  - info:
      name: Call Tool
      type: http
    http:
      method: POST
      url: https://api.klavis.ai/api/mcp-server/call-tool
      body:
        type: json
        data: '{}'
    docs: Invoke a tool on a specific remote MCP server. Used by clients to perform function calls without writing per-tool
      MCP code.
- info:
    name: Sandbox
    type: folder
  items:
  - info:
      name: Acquire Sandbox
      type: http
    http:
      method: POST
      url: https://api.klavis.ai/api/sandbox/create
      body:
        type: json
        data: '{}'
    docs: Acquire an idle sandbox instance for a specific MCP server with optional benchmark configuration.
- info:
    name: Users
    type: folder
  items:
  - info:
      name: Get User Information
      type: http
    http:
      method: GET
      url: https://api.klavis.ai/api/user/:user_id
      params:
      - name: user_id
        value: ''
        type: path
        description: User identifier
    docs: Retrieve user metadata by user identifier.
bundled: true

Klavis AI

APIs

Klavis MCP Platform API

Collections

Klavis AI MCP Platform API

Pricing Plans

Klavis Plans Pricing

Rate Limits

Klavis Rate Limits

FinOps

Klavis Finops

Resources

Sources