Implement LangGraph integration example

[adam2go]

Summary

Add a complete working example showing how to use Tilo with LangGraph.

LangGraph is the most widely used stateful agent framework (from the LangChain team). Many developers build multi-step agents with LangGraph and want to render the output as interactive UI.

What to build

File: examples/integrations/langgraph_example.py

The example should show:

1. A simple LangGraph graph (at least 2 nodes)
2. Using TiloCallbackHandler to capture the final output
3. Converting the graph output to a Tilo AIP spec via langchain_result_to_spec
4. Printing the block types produced

Acceptance criteria

• Works with pip install tilo langchain-openai langgraph
• Runs without errors when OPENAI_API_KEY is set
• Prints a deterministic placeholder when no API key is present
• Added to examples/integrations/README.md table

Resources

• examples/integrations/langchain_example.py — existing LangChain example to follow
• backend/tilo/adapters/langchain.py — the adapter to use
• LangGraph docs: https://langchain-ai.github.io/langgraph/