Skip to main content

Examples / Cookbook

Start with the complete multi-SDK implementation first, then use cookbooks for focused patterns. Full implementation page: Full-fledged support agent Then apply recipes in this order:
  1. 01-quickstart-support-agent.py -> Support agent memory loop
  2. 07-multi-agent-system.py -> Multi-agent shared state
  3. 08-event-driven-agents.py -> Building event-driven agents
  4. 09-branch-what-if.py -> Branching what-if simulations
  5. 10-stateful-task-tree.py -> Stateful task trees

Framework integration examples

Complete example apps with real LLM calls showing MuBit integrated with popular agent frameworks:
ExampleFrameworkWhat it shows
Support Ticket TriageCrewAI3-agent crew with handoffs, checkpoints, and outcome recording
Code Review PipelineLangGraphStateGraph with conditional reviewer loop and MuBit store
Research AssistantLangChainMulti-turn conversation with cross-session memory
Travel PlannerGoogle ADKSequentialAgent with Gemini tool calling
FAQ BotVercel AI SDKMiddleware auto-injection with knowledge-base tool
See each example’s directory for setup and run instructions.

Failure modes and troubleshooting

SymptomRoot causeFix
Recipes work alone but fail togetherRun identity not carried between stagesKeep deterministic run IDs by role/task
Progression feels disconnectedCookbook file order skippedFollow stage order and reuse the same project folder

Next steps