Testing a chatbot agent in Claude Code¶
This tutorial has been merged into the unified Chatbot agent end-to-end test.
The merged tutorial covers both test paths:
- Path A — Portal Chat tab: production-fidelity test with your configured LLM (costs API credits per turn).
- Path B — Claude Code (MCP): free, fast iteration loop where Claude Code plays the chatbot. The original content of this page lives there.
Pick the path that matches your goal at the top of the merged tutorial, then follow the tabbed sections in Parts 4 and 5. The agent and conversation design (Parts 1–3) are shared between both paths.