Alteryx MCP Server and Alteryx Skills
Connect agents to Alteryx One with MCP Server and Skills. They help teams scale agent-driven work by making it traceable, repeatable, and trustworthy.
Connect agents to Alteryx One with MCP Server and Skills. They help teams scale agent-driven work by making it traceable, repeatable, and trustworthy.
Connect supported AI applications to Alteryx One to build and operate Alteryx workflows, run them on demand, pull governed answers, and automate actions with confidence.
Use tools such as Claude Code, Gemini CLI, and Codex to create, inspect, test, and refine Alteryx workflows and analytics assets.
By connecting to Alteryx, AI can understand your business and take action without needing to open the Alteryx UI.
Connect your AI assistant to other applications, and use Alteryx to automate the actions, not just surface answers.
Requests are authenticated and executed according to the applicable Alteryx workspace, role, and asset permissions, helping administrators retain control over what each connection can access and run.
Connect through the Alteryx MCP Server and Skills, and every AI tool returns real ROI: answers you can trust, lower token costs, and faster time to insights.
Give AI agents access to durable Alteryx workflows instead of rebuilding the same transformations and calculations for every application.
Perform data-intensive and deterministic processing in Alteryx, then return focused results to the language model, reducing the amount of data the model needs to process and number of tokens to use.
Apply established authentication, permissions, execution controls, and auditability when external AI tools use approved Alteryx capabilities.
With more context to work with, AI has less to infer, so the answers you get come from your actual data and business logic instead of a guess.
Organizations use the Alteryx MCP Server and Skills to let any AI tool work with governed Alteryx data and workflows without compromising security.
95%
reduced token usage
NextWave reported that combining deterministic processing in Alteryx with Claude through MCP reduced token usage by more than 95% while preserving the quality of the resulting business insights.
Connecting to Alteryx One from outside its UI looks like this:
Install the MCP Server and Skills in a supported assistant or agent.
A user logs into their workspace and access controls and permissions are respected.
Use the MCP Server and Skills to create, analyze, and invoke trusted business logic in Alteryx.
Underpin your business processes with visible, understandable, repeatable, auditable Alteryx workflows that prepare and generate AI-ready data and context needed for accurate, cost-effective AI.
Alteryx MCP Server is a governed connection between Alteryx One and supported MCP-compatible AI applications. It allows those applications to discover and invoke approved Alteryx capabilities using the open Model Context Protocol.
Alteryx Skills are reusable instruction plugins that help supported coding agents perform Alteryx-specific tasks consistently. They provide procedural guidance for activities such as building, testing, and operating Alteryx workflows.
MCP Server gives an AI agent access to callable Alteryx capabilities. Skills provide the Alteryx-specific knowledge and procedures that help a coding agent use those capabilities effectively.
Organizations can bring trusted Alteryx data and analytics into their preferred AI experiences, reuse established workflows and business logic, and maintain control over access and execution.
Alteryx supports selected MCP-compatible clients, including those from OpenAI, Microsoft, Anthropic, Google and others. Supported clients and available capabilities may vary by deployment, authentication method, and product release.
Builders can build, test, and operate workflows faster from inside a coding agent like Claude Code, Gemini CLI, or Codex.
Viewers use the Alteryx MCP Server through everyday AI assistants like Claude, ChatGPT, or Gemini to ask questions and get governed, trusted answers pulled from workflows the team already built, without needing to build anything themselves.
Admins manage which workflows, datasets, and toolsets are exposed to AI applications in the first place, so every answer or action an agent produces stays inside the same permissions and governance controls already in place for that user.