diff --git a/trustgraph-base/trustgraph/api/async_socket_client.py b/trustgraph-base/trustgraph/api/async_socket_client.py
index b68a69d5..7c2a5aab 100644
--- a/trustgraph-base/trustgraph/api/async_socket_client.py
+++ b/trustgraph-base/trustgraph/api/async_socket_client.py
@@ -130,18 +130,26 @@ class AsyncSocketClient:
                 content=resp.get("content", ""),
                 end_of_message=resp.get("end_of_message", False)
             )
-        elif chunk_type == "final-answer":
+        elif chunk_type == "answer" or chunk_type == "final-answer":
             return AgentAnswer(
                 content=resp.get("content", ""),
                 end_of_message=resp.get("end_of_message", False),
                 end_of_dialog=resp.get("end_of_dialog", False)
             )
+        elif chunk_type == "action":
+            # Agent action chunks - treat as thoughts for display purposes
+            return AgentThought(
+                content=resp.get("content", ""),
+                end_of_message=resp.get("end_of_message", False)
+            )
         else:
             # RAG-style chunk (or generic chunk)
+            # Text-completion uses "response" field, RAG uses "chunk" field, Prompt uses "text" field
+            content = resp.get("response", resp.get("chunk", resp.get("text", "")))
             return RAGChunk(
-                content=resp.get("chunk", ""),
+                content=content,
                 end_of_stream=resp.get("end_of_stream", False),
-                error=resp.get("error")
+                error=None  # Errors are always thrown, never stored
             )
 
     async def aclose(self):
diff --git a/trustgraph-base/trustgraph/api/socket_client.py b/trustgraph-base/trustgraph/api/socket_client.py
index 09e937e6..c0fd9cd9 100644
--- a/trustgraph-base/trustgraph/api/socket_client.py
+++ b/trustgraph-base/trustgraph/api/socket_client.py
@@ -202,12 +202,18 @@ class SocketClient:
                 content=resp.get("content", ""),
                 end_of_message=resp.get("end_of_message", False)
             )
-        elif chunk_type == "final-answer":
+        elif chunk_type == "answer" or chunk_type == "final-answer":
             return AgentAnswer(
                 content=resp.get("content", ""),
                 end_of_message=resp.get("end_of_message", False),
                 end_of_dialog=resp.get("end_of_dialog", False)
             )
+        elif chunk_type == "action":
+            # Agent action chunks - treat as thoughts for display purposes
+            return AgentThought(
+                content=resp.get("content", ""),
+                end_of_message=resp.get("end_of_message", False)
+            )
         else:
             # RAG-style chunk (or generic chunk)
             # Text-completion uses "response" field, RAG uses "chunk" field, Prompt uses "text" field
diff --git a/trustgraph-base/trustgraph/api/types.py b/trustgraph-base/trustgraph/api/types.py
index 016f2c7a..bba566a6 100644
--- a/trustgraph-base/trustgraph/api/types.py
+++ b/trustgraph-base/trustgraph/api/types.py
@@ -79,5 +79,6 @@ class AgentAnswer(StreamingChunk):
 @dataclasses.dataclass
 class RAGChunk(StreamingChunk):
     """RAG streaming chunk"""
+    chunk_type: str = "rag"
     end_of_stream: bool = False
     error: Optional[Dict[str, str]] = None