generated: '2026-06-20'
method: searched
status: published
source: https://docs.airbyte.com/developers/mcp-servers
servers:
- name: airbyte-knowledge-mcp
  title: Airbyte Knowledge MCP Server
  transport: http
  url: https://airbyte.mcp.kapa.ai
  install: claude mcp add --transport http airbyte-docs https://airbyte.mcp.kapa.ai
  auth: Google account (rate-limited; 40 req/hour, 200 req/day per user)
  description: Semantic search across Airbyte docs, website, OpenAPI specs, YouTube videos, and GitHub
    issues/discussions/PRs. Helps agents build connectors, use PyAirbyte, configure the API, and troubleshoot.
  docs: https://docs.airbyte.com/community/mcp-servers/airbyte-knowledge-mcp
- name: pyairbyte-mcp
  title: PyAirbyte MCP Server
  transport: stdio
  install: uvx --python=3.11 --from=airbyte@latest airbyte-mcp
  package: airbyte (PyPI)
  status: experimental
  description: Run and manage Airbyte connectors locally through PyAirbyte. Tools list available connectors,
    fetch connector info, read dotenv secrets, and inspect Airbyte Cloud workspace connections and destinations.
  docs: https://docs.airbyte.com/developers/mcp-servers/pyairbyte-mcp
- name: airbyte-agent-mcp
  title: Agent MCP (Airbyte Agents)
  transport: http
  description: Connects MCP-capable agents (Claude, Cursor, VS Code, ChatGPT, Codex) to business data
    through Airbyte Agents' 600+ connectors. Credentials stay in the MCP server; the LLM only sees env
    var names.
  docs: https://docs.airbyte.com/ai-agents
- name: airbyte-replication-mcp
  title: Airbyte Replication MCP Server
  transport: stdio
  description: Manage Airbyte data replication through MCP clients - list connectors, validate configurations,
    and run sync operations.
  docs: https://docs.airbyte.com/community/mcp-servers/pyairbyte-mcp
deployment:
  mode: remote
  endpoint: https://airbyte.mcp.kapa.ai
  verified: probed
  probe: gated
  checked: '2026-08-12'
  source: catalog MCP census