Checkpoint from Copilot CLI for coding agent session

plans/castchat.md

@@ -0,0 +1,76 @@
+# castchat
+
+### Overview
+
+With podcast episodes transcribed and the transcription data indexed,
+we want an interactive REPL to ask questions of the indexed data.
+
+We use chromadb and PydanticAI to implement the tool enabled and RAG
+based chat for cli application named `castchat`. We want to be able to
+invoke this cli as a subcommand of retrocast, e.g. `retrocast
+castchat`.
+
+Below is an example of creating a tool to integrate with Pydantic’s 
+`clai` agentic REPL toolkit.
+
+```python
+from pydantic_ai import Agent
+from pydantic_ai.models.anthropic import AnthropicModel
+import chromadb
+
+# Setup ChromaDB
+chroma_client = chromadb.Client()
+collection = chroma_client.get_or_create_collection("docs")
+
+# Add sample documents (do this once)
+collection.add(
+    documents=[
+        "Our refund policy: Full refunds within 30 days of purchase",
+        "Customer support: Email support@example.com or call 1-800-HELP",
+        "Shipping: Free shipping on orders over $50, 5-7 business days"
+    ],
+    ids=["doc1", "doc2", "doc3"]
+)
+
+# Create agent
+model = AnthropicModel('claude-sonnet-4-20250514')
+agent = Agent(model=model)
+
+@agent.tool
+def search_docs(query: str) -> str:
+    """Search internal documentation for relevant information.
+
+    Args:
+        query: The search query text
+    """
+    results = collection.query(
+        query_texts=[query],
+        n_results=3
+    )
+
+    if results["documents"] and results["documents"][0]:
+        docs = "\n\n".join(results["documents"][0])
+        return f"Found relevant documentation:\n{docs}"
+    return "No relevant documents found"
+```
+
+Here’s an example of firing up `clai` to interact using that tool
+
+````bash
+# Start interactive session
+clai --agent search_agent.py:agent
+
+# Or specify model explicitly
+clai --agent search_agent.py:agent --model anthropic:claude-sonnet-4-20250514
+```
+
+**3. Example conversation:**
+```
+You: What's the refund policy?
+Agent: [calls search_docs tool, then responds with the policy]
+
+You: How do I contact support?
+Agent: [searches and provides contact info]
+````
+
+