adding getter

crates/potato_type/src/anthropic/v1/request.rs

@@ -1225,6 +1225,7 @@ impl MessageParam {
     }
 
     // Return the text content from the first content part that is text
+    #[getter]
     pub fn text(&self) -> String {
         self.content
             .iter()

crates/potato_type/src/google/v1/generate/request.rs

@@ -1382,6 +1382,7 @@ impl GeminiContent {
     }
 
     // helper method for returning firs text part content
+    #[getter]
     pub fn text(&self) -> String {
         self.parts
             .iter()

crates/potato_type/src/openai/v1/chat/request.rs

@@ -328,6 +328,7 @@ impl ChatMessage {
     }
 
     // Return the text content from the first content part that is text
+    #[getter]
     pub fn text(&self) -> String {
         self.content
             .iter()

py-potato/examples/anthropic/structured_output.py

@@ -38,7 +38,7 @@ class StructuredTaskOutput(BaseModel):
 agent = Agent(Provider.Anthropic)
 
 if __name__ == "__main__":
-    result: StructuredTaskOutput = agent.execute_prompt(
+    result = agent.execute_prompt(
         prompt=prompt,
         output_type=StructuredTaskOutput,
     ).structured_output

py-potato/examples/google/query_reformulation.py

@@ -5,8 +5,13 @@
 # 2. Evaluates the quality of the reformulated query.
 # 3. Returns a score and reason for how well the reformulation improves the query.
 
-from potato_head import Agent, Prompt, Provider, Score, Task, Workflow
-from potato_head.google import GeminiSettings, GenerationConfig, GeminiThinkingConfig
+from potato_head import Agent, Prompt, Provider, Score, Task, Workflow, AgentResponse
+from potato_head.google import (
+    GeminiSettings,
+    GenerationConfig,
+    GeminiThinkingConfig,
+    GenerateContentResponse,
+)
 from potato_head.logging import LoggingConfig, LogLevel, RustyLogger
 
 RustyLogger.setup_logging(LoggingConfig(log_level=LogLevel.Debug))
@@ -112,7 +117,9 @@ def create_workflow():
 
     agent = Agent(Provider.Gemini)
     user_query = "How do I find good post-hardcore bands?"
-    response = agent.execute_prompt(prompt=prompt.bind(user_query=user_query))
+    response: AgentResponse[str, GenerateContentResponse] = agent.execute_prompt(
+        prompt=prompt.bind(user_query=user_query)
+    )
 
     workflow = create_workflow()
 

py-potato/examples/openai/structured_output.py

@@ -35,7 +35,7 @@ class StructuredTaskOutput(BaseModel):
 agent = Agent(Provider.OpenAI)
 
 if __name__ == "__main__":
-    result: StructuredTaskOutput = agent.execute_prompt(
+    result = agent.execute_prompt(
         prompt=prompt,
         output_type=StructuredTaskOutput,
     ).structured_output

py-potato/python/potato_head/_potato_head.pyi

@@ -714,8 +714,9 @@ OutT = TypeVar(
     "OutT",
     default=str,
 )
+RespT = TypeVar("RespT", default=_ResponseType)
 
-class AgentResponse(Generic[OutT]):
+class AgentResponse(Generic[OutT, RespT]):
     """Agent response generic over OutputDataT.
 
     The structured_output property returns OutputDataT type.
@@ -731,7 +732,7 @@ class AgentResponse(Generic[OutT]):
         """The ID of the agent response."""
 
     @property
-    def response(self) -> _ResponseType:
+    def response(self) -> RespT:
         """The response of the agent."""
 
     @property
@@ -847,7 +848,7 @@ class Agent:
         self,
         task: Task,
         output_type: type[OutT] | None = None,
-    ) -> AgentResponse[OutT]:
+    ) -> AgentResponse[OutT, _ResponseType]:
         """Execute a task.
 
         Args:
@@ -857,15 +858,16 @@ class Agent:
                 The output type to use for the task.
 
         Returns:
-            AgentResponse[OutT]:
-                The response from the agent after executing the task.
+            AgentResponse[OutT, _ResponseType]:
+                The response from the agent. For type-safe response access,
+                annotate the return value with the specific response type.
         """
 
     def execute_prompt(
         self,
         prompt: Prompt,
         output_type: type[OutT] | None = None,
-    ) -> AgentResponse[OutT]:
+    ) -> AgentResponse[OutT, _ResponseType]:
         """Execute a prompt.
 
         Args:
@@ -875,8 +877,9 @@ class Agent:
                 The output type to use for the task.
 
         Returns:
-            AgentResponse:
-                The response from the agent after executing the task.
+            AgentResponse[OutT, _ResponseType]:
+                The response from the agent. For type-safe response access,
+                annotate the return value with the specific response type.
         """
 
     @property
@@ -2540,6 +2543,7 @@ class ChatMessage:
             TypeError: If content format is invalid
         """
 
+    @property
     def text(self) -> str:
         """Get the text content of the first part, if available. Returns
         an empty string if the first part is not text.
@@ -5322,6 +5326,7 @@ class GeminiContent:
                 Role of the message sender (e.g., "user", "model", "function")
         """
 
+    @property
     def text(self) -> str:
         """Get the text content of the first part, if available. Returns
         an empty string if the first part is not text.
@@ -8648,6 +8653,7 @@ class MessageParam:
                 Message role ("user" or "assistant")
         """
 
+    @property
     def text(self) -> str:
         """Get the text content of the first part, if available. Returns
         an empty string if the first part is not text.

py-potato/tests/prompt/test_anthropic.py

@@ -30,8 +30,8 @@ def test_prompt():
         system_instructions="system_prompt",
     )
 
-    assert prompt.anthropic_message.text() == "My prompt"
-    assert prompt.system_instructions[0].text() == "system_prompt"
+    assert prompt.anthropic_message.text == "My prompt"
+    assert prompt.system_instructions[0].text == "system_prompt"
 
     # test string message
     prompt = Prompt(
@@ -76,8 +76,8 @@ def test_bind_prompt():
     )
 
     bound_prompt = prompt.bind("variable1", "world").bind("variable2", "Foo")
-    assert bound_prompt.messages[0].content[0].text == "Hello world"
-    assert bound_prompt.messages[1].content[0].text == "This is Foo"
+    assert bound_prompt.messages[0].text == "Hello world"
+    assert bound_prompt.messages[1].text == "This is Foo"
 
 
 def test_prompt_structured_output():

py-potato/tests/prompt/test_gemini.py

@@ -79,8 +79,8 @@ def test_prompt():
 
 def test_bind_prompt():
     prompt = Prompt(
-        model="gemini-3.0-flash",
         provider="gemini",
+        model="gemini-3.0-flash",
         messages=[
             "Hello ${variable1}",
             "This is ${variable2}",
@@ -90,7 +90,8 @@ def test_bind_prompt():
     )
 
     bound_prompt = prompt.bind("variable1", "world").bind("variable2", "Foo")
-    assert bound_prompt.messages[0].parts[0].data == "Hello world"
+    messages = bound_prompt.messages
+    assert messages[0].parts[0].data == "Hello world"
     assert bound_prompt.messages[1].parts[0].data == "This is Foo"
 
 

py-potato/tests/prompt/test_openai.py

@@ -98,9 +98,11 @@ def test_bind_prompt():
         ],
         system_instructions="system_prompt",
     )
-    bound_prompt = prompt.bind("variable1", "world").bind("variable2", "Foo")
-    assert bound_prompt.openai_messages[0].content[0].text == "Hello world"
-    assert bound_prompt.openai_messages[1].content[0].text == "This is Foo"
+    kwargs = {"variable1": "world", "variable2": "Foo"}
+    bound_prompt = prompt.bind(**kwargs)
+
+    assert bound_prompt.openai_messages[0].text == "Hello world"
+    assert bound_prompt.openai_messages[1].text == "This is Foo"
 
     # testing binding with kwargs
     bound_prompt = prompt.bind(variable1="world")

py-potato/tests/test_workflow.py

@@ -33,23 +33,23 @@ class Prompts:
 def test_simple_workflow(prompt_step1: Prompt):
     agent = PydanticAgent(
         prompt_step1.model_identifier,
-        system_prompt=prompt_step1.system_instructions[0].text(),
+        system_prompt=prompt_step1.system_instructions[0].text,
     )
 
     with agent.override(model=TestModel()):
-        agent.run_sync(prompt_step1.message.text())
+        agent.run_sync(prompt_step1.message.text)
 
 
 def test_simple_dep_workflow(prompt_step1: Prompt, prompt_step2: Prompt):
     agent = PydanticAgent(
         prompt_step1.model_identifier,
-        system_prompt=prompt_step1.system_instructions[0].text(),
+        system_prompt=prompt_step1.system_instructions[0].text,
         deps_type=Prompts,
     )
 
     @agent.system_prompt
     def get_system_instruction(ctx: RunContext[Prompts]) -> str:
-        return ctx.deps.prompt_step1.system_instructions[0].text()
+        return ctx.deps.prompt_step1.system_instructions[0].text
 
     with agent.override(model=TestModel()):
         agent.run_sync(
@@ -64,7 +64,7 @@ def get_system_instruction(ctx: RunContext[Prompts]) -> str:
 def test_binding_workflow(prompt_step1: Prompt, prompt_step2: Prompt):
     agent = PydanticAgent(
         "openai:gpt-4o",
-        system_prompt=prompt_step1.system_instructions[0].text(),
+        system_prompt=prompt_step1.system_instructions[0].text,
         deps_type=Prompts,
     )