You're in your terminal — Claude Code, Codex, whichever your team uses. Ask the agent for a DataWeave transformation and you'll get something, plausible-looking code that may or may not survive contact with real data, null fields, or a target schema the agent hasn't fully internalized.
CurieTech AI is integration intelligence built specifically for Integration teams. If Claude Code or Codex is where your work happens, we're now a tool the agent can reach for, without leaving the terminal.
With the Curie MCP server connected, Claude Code stops guessing at DataWeave and hands the job to a model that already knows it, the fastest path from prompt to a transformation that runs.
And it's not just JSON and XML. CurieTech AI handles JSON, XML, YAML, CSV, plain text, EDI, EDIFACT, and HL7 so the same workflow covers the messier integration formats that usually need their own tooling.
What you can do with Curie MCP
A few real prompts to try, right from your terminal:
- “Use CurieTech AI to generate a DataWeave transformation for input.json that maps customer details to the schema in account-schema.json.”
- “Use CurieTech AI to update transform.dwl so it handles missing or null fields. Use input.json and expected-output.json as examples.”
- “Use CurieTech AI to refactor this transformation for readability without changing its output.”
Why this matters for teams
Integration work has long lived across too many tools. Pulling CurieTech AI into the same loop Claude Code is already running means engineers stay in flow, reviewers see the reasoning in the same place as the diff, and onboarding a new developer stops requiring a tour of multiple surfaces.
And because it’s exposed through MCP, CurieTech AI isn’t a separate stop the engineer has to remember; it’s a capability the agent can call when the task calls for it - pulling in project files, applying the result back to the codebase, and keeping the loop tight.
If your team is on our VS Code or Anypoint Studio plugins, nothing changes - they work the way they always have. This is purely additive: where your workflow has moved into Claude Code, CurieTech AI now goes with you.
What’s next
This is our first MCP surface, and we’re starting with DataWeave. From here, we’ll keep expanding what’s available through the CurieTech AI MCP server and bring it to more of the agentic environment developers use. The goal stays the same: meet teams where they already build, without asking them to change how they work.
Get connected — under two minutes
Create a CurieTech account, if you don’t have one. Grab an API key from platform.curietech.ai under Settings → API Keys, then add the server to Claude Code:
claude mcp add --transport http curie-dataweave https://platform.curietech.ai/mcp/ --header "Authorization: Bearer <API_KEY>"
Run /mcp to confirm curie-dataweave shows as connected. Then open a project and try one of the prompts above.
