diff --git a/examples/01_basic_agent.py b/examples/01_basic_agent.py
index 93794cc..8d9f266 100644
--- a/examples/01_basic_agent.py
+++ b/examples/01_basic_agent.py
@@ -5,9 +5,11 @@
 1. Basic agent creation with input/output models
 2. Field descriptions and examples
 3. Cost and latency tracking
+4. How to fetch and analyze completions after a run
 """
 
 import asyncio
+from typing import Any
 
 from pydantic import BaseModel, Field
 
@@ -17,6 +19,7 @@
 
 class CityInput(BaseModel):
     """Input model for the city-to-capital agent."""
+
     city: str = Field(
         description="The name of the city for which to find the country's capital",
         examples=["Paris", "New York", "Tokyo"],
@@ -25,6 +28,7 @@ class CityInput(BaseModel):
 
 class CapitalOutput(BaseModel):
     """Output model containing information about the capital city."""
+
     country: str = Field(
         description="The country where the input city is located",
         examples=["France", "United States", "Japan"],
@@ -57,6 +61,21 @@ async def get_capital_info(city_input: CityInput) -> Run[CapitalOutput]:
     ...
 
 
+async def display_completions(run: Run[Any]) -> None:
+    """Helper function to display completions for a run."""
+    try:
+        completions = await run.fetch_completions()
+
+        for completion in completions:
+            print("\n--- Completion Details ---")
+
+            # Use model_dump_json for clean serialization
+            completion_json = completion.model_dump_json(indent=2)
+            print(completion_json)
+    except (ValueError, workflowai.WorkflowAIError) as e:
+        print(f"Error: {e}")
+
+
 async def main():
     # Example 1: Basic usage with Paris
     print("\nExample 1: Basic usage with Paris")
@@ -70,6 +89,9 @@ async def main():
     run = await get_capital_info.run(CityInput(city="Tokyo"))
     print(run)
 
+    # Fetch and display completions for the Tokyo example
+    await display_completions(run)
+
 
 if __name__ == "__main__":
     asyncio.run(main())
diff --git a/examples/02_agent_with_tools.py b/examples/02_agent_with_tools.py
index 85289e5..9bfb008 100644
--- a/examples/02_agent_with_tools.py
+++ b/examples/02_agent_with_tools.py
@@ -62,7 +62,7 @@ class HistoricalEventOutput(BaseModel):
 
 @workflowai.agent(
     id="historical-event-analyzer",
-    model=Model.GEMINI_1_5_FLASH_LATEST,
+    model=Model.GEMINI_2_0_FLASH_LATEST,
     tools=[get_current_date, calculate_days_between],
 )
 async def analyze_historical_event(event_input: HistoricalEventInput) -> HistoricalEventOutput:
diff --git a/examples/03_caching.py b/examples/03_caching.py
index e55d85c..5e3bd7a 100644
--- a/examples/03_caching.py
+++ b/examples/03_caching.py
@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 # Import CacheUsage type
 CacheUsage = Literal["auto", "always", "never"]
@@ -54,7 +54,7 @@ class SOAPNote(BaseModel):
     id="soap-extractor",
     model=Model.LLAMA_3_3_70B,
 )
-async def extract_soap_notes(soap_input: SOAPInput) -> Run[SOAPNote]:
+async def extract_soap_notes(soap_input: SOAPInput) -> SOAPNote:
     """
     Extract SOAP notes from a medical consultation transcript.
 
@@ -91,7 +91,7 @@ async def demonstrate_caching(transcript: str):
     for cache_option in cache_options:
         start_time = time.time()
 
-        run = await extract_soap_notes(
+        run = await extract_soap_notes.run(
             SOAPInput(transcript=transcript),
             use_cache=cache_option,
         )
diff --git a/examples/04_audio_classifier_agent.py b/examples/04_audio_classifier_agent.py
index 8bf5a19..fb36246 100644
--- a/examples/04_audio_classifier_agent.py
+++ b/examples/04_audio_classifier_agent.py
@@ -13,13 +13,13 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
-from workflowai.fields import File
+from workflowai import Model
+from workflowai.fields import Audio
 
 
 class AudioInput(BaseModel):
     """Input containing the audio file to analyze."""
-    audio: File = Field(
+    audio: Audio = Field(
         description="The audio recording to analyze for spam/robocall detection",
     )
 
@@ -67,7 +67,7 @@ class AudioClassification(BaseModel):
     id="audio-spam-detector",
     model=Model.GEMINI_1_5_FLASH_LATEST,
 )
-async def classify_audio(audio_input: AudioInput) -> Run[AudioClassification]:
+async def classify_audio(audio_input: AudioInput) -> AudioClassification:
     """
     Analyze the audio recording to determine if it's a spam/robocall.
 
@@ -108,18 +108,18 @@ async def main():
     with open(audio_path, "rb") as f:  # noqa: ASYNC230
         audio_data = f.read()
 
-    audio = File(
+    audio = Audio(
         content_type="audio/mp3",
         data=base64.b64encode(audio_data).decode(),
     )
 
     # Example 2: Using a URL instead of base64 (commented out)
-    # audio = File(
+    # audio = Audio(
     #     url="https://example.com/audio/call.mp3"
     # )
 
     # Classify the audio
-    run = await classify_audio(AudioInput(audio=audio))
+    run = await classify_audio.run(AudioInput(audio=audio))
 
     # Print results including cost and latency information
     print(run)
diff --git a/examples/05_browser_text_uptime_agent.py b/examples/05_browser_text_uptime_agent.py
index 4e47624..a5d0243 100644
--- a/examples/05_browser_text_uptime_agent.py
+++ b/examples/05_browser_text_uptime_agent.py
@@ -11,7 +11,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class UptimeInput(BaseModel):
@@ -40,7 +40,7 @@ class UptimeOutput(BaseModel):
     id="uptime-checker",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def check_uptime(uptime_input: UptimeInput, use_cache: str = "never") -> Run[UptimeOutput]:
+async def check_uptime(uptime_input: UptimeInput, use_cache: str = "never") -> UptimeOutput:
     """
     Fetch and analyze uptime data from an API status page.
     Use @browser-text to get the page content.
@@ -68,7 +68,7 @@ async def main():
     print("-" * 50)
 
     # Get uptime data with caching disabled
-    run = await check_uptime(uptime_input, use_cache="never")
+    run = await check_uptime.run(uptime_input, use_cache="never")
 
     # Print the run
     print(run)
diff --git a/examples/06_streaming_summary.py b/examples/06_streaming_summary.py
index 557cb3b..23617bc 100644
--- a/examples/06_streaming_summary.py
+++ b/examples/06_streaming_summary.py
@@ -6,12 +6,11 @@
 """
 
 import asyncio
-from collections.abc import AsyncIterator
 
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class TranslationInput(BaseModel):
@@ -30,7 +29,7 @@ class TranslationOutput(BaseModel):
 
 
 @workflowai.agent(id="french-translator", model=Model.CLAUDE_3_5_SONNET_LATEST)
-def translate_to_english(_: TranslationInput) -> AsyncIterator[Run[TranslationOutput]]:
+async def translate_to_english(_: TranslationInput) -> TranslationOutput:
     """
     Translate French text into natural, fluent English.
 
@@ -74,7 +73,7 @@ async def main():
     # This ensures we can see the streaming effect in the example
     # Otherwise, subsequent runs would return the cached result instantly,
     # making it hard to observe the incremental streaming behavior
-    async for chunk in translate_to_english(TranslationInput(text=french_text), use_cache="never"):
+    async for chunk in translate_to_english.stream(TranslationInput(text=french_text), use_cache="never"):
         print(f"--- Translation Progress (Chunk {chunk_num}) ---")
         print(chunk.output.translation)
         print("-" * 50)
diff --git a/examples/08_pdf_agent.py b/examples/08_pdf_agent.py
index f2e849d..4cce803 100644
--- a/examples/08_pdf_agent.py
+++ b/examples/08_pdf_agent.py
@@ -5,13 +5,13 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Run, WorkflowAIError
+from workflowai import WorkflowAIError
 from workflowai.core.domain.model import Model
-from workflowai.fields import File
+from workflowai.fields import PDF
 
 
 class PDFQuestionInput(BaseModel):
-    pdf: File = Field(description="The PDF document to analyze")
+    pdf: PDF = Field(description="The PDF document to analyze")
     question: str = Field(description="The question to answer about the PDF content")
 
 
@@ -21,7 +21,7 @@ class PDFAnswerOutput(BaseModel):
 
 
 @workflowai.agent(id="pdf-answer", model=Model.CLAUDE_3_5_SONNET_LATEST)
-async def answer_pdf_question(_: PDFQuestionInput) -> Run[PDFAnswerOutput]:
+async def answer_pdf_question(_: PDFQuestionInput) -> PDFAnswerOutput:
     """
     Analyze the provided PDF document and answer the given question.
     Provide a clear and concise answer based on the content found in the PDF.
@@ -49,13 +49,13 @@ async def run_pdf_answer():
 
         content = base64.b64encode(pdf_file.read()).decode("utf-8")
 
-    pdf = File(content_type="application/pdf", data=content)
+    pdf = PDF(content_type="application/pdf", data=content)
     # Could also pass the content via url
-    # pdf = File(url="https://example.com/sample.pdf")
+    # pdf = PDF(url="https://example.com/sample.pdf")
     question = "How many stocks were sold? What is the total amount in USD?"
 
     try:
-        agent_run = await answer_pdf_question(
+        run = await answer_pdf_question.run(
             PDFQuestionInput(pdf=pdf, question=question),
             use_cache="auto",
         )
@@ -63,9 +63,7 @@ async def run_pdf_answer():
         print(f"Failed to run task. Code: {e.error.code}. Message: {e.error.message}")
         return
 
-    print("\n--------\nAgent output:\n", agent_run.output, "\n--------\n")
-    print(f"Cost: ${agent_run.cost_usd:.10f}")
-    print(f"Latency: {agent_run.duration_seconds:.2f}s")
+    print(run)
 
 
 if __name__ == "__main__":
diff --git a/examples/09_reply.py b/examples/09_reply.py
index f5bdfed..9fccca9 100644
--- a/examples/09_reply.py
+++ b/examples/09_reply.py
@@ -20,7 +20,7 @@
 from pydantic import BaseModel, Field  # pyright: ignore [reportUnknownVariableType]
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class NameExtractionInput(BaseModel):
@@ -43,7 +43,7 @@ class NameExtractionOutput(BaseModel):
 
 
 @workflowai.agent(id="name-extractor", model=Model.GPT_4O_MINI_LATEST)
-async def extract_name(_: NameExtractionInput) -> Run[NameExtractionOutput]:
+async def extract_name(_: NameExtractionInput) -> NameExtractionOutput:
     """
     Extract a person's first and last name from a sentence.
     Be precise and consider cultural variations in name formats.
@@ -64,7 +64,7 @@ async def main():
         print(f"\nProcessing: {sentence}")
 
         # Initial extraction
-        run = await extract_name(NameExtractionInput(sentence=sentence))
+        run = await extract_name.run(NameExtractionInput(sentence=sentence))
 
         print(f"Extracted: {run.output.first_name} {run.output.last_name}")
 
diff --git a/examples/10_calendar_event_extraction.py b/examples/10_calendar_event_extraction.py
index be1c3fe..4675802 100644
--- a/examples/10_calendar_event_extraction.py
+++ b/examples/10_calendar_event_extraction.py
@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 from workflowai.fields import File
 
 
@@ -79,7 +79,7 @@ class CalendarEventOutput(BaseModel):
     id="calendar-event-extractor",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def extract_calendar_event_from_email(email_input: EmailInput) -> Run[CalendarEventOutput]:
+async def extract_calendar_event_from_email(email_input: EmailInput) -> CalendarEventOutput:
     """
     Extract calendar event details from email content.
 
@@ -112,7 +112,7 @@ async def extract_calendar_event_from_email(email_input: EmailInput) -> Run[Cale
     id="calendar-event-extractor",
     model=Model.GPT_4O_MINI_LATEST,
 )
-async def extract_calendar_event_from_image(image_input: ImageInput) -> Run[CalendarEventOutput]:
+async def extract_calendar_event_from_image(image_input: ImageInput) -> CalendarEventOutput:
     """
     Extract calendar event details from an event poster or flyer image.
 
@@ -160,7 +160,7 @@ async def main():
         """,
     )
 
-    run = await extract_calendar_event_from_email(email1)
+    run = await extract_calendar_event_from_email.run(email1)
     print(run)
 
     # Example 2: Virtual meeting with more details
@@ -238,7 +238,7 @@ async def main():
     )
 
     try:
-        run = await extract_calendar_event_from_email(email4)
+        run = await extract_calendar_event_from_email.run(email4)
         print(run)
     except workflowai.WorkflowAIError as e:
         print(f"As expected, no calendar event found: {e!s}")
diff --git a/examples/12_contextual_retrieval.py b/examples/12_contextual_retrieval.py
index aacd7f0..796a411 100644
--- a/examples/12_contextual_retrieval.py
+++ b/examples/12_contextual_retrieval.py
@@ -9,7 +9,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class ContextGeneratorInput(BaseModel):
@@ -37,7 +37,7 @@ class ContextGeneratorOutput(BaseModel):
     id="context-generator",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def generate_chunk_context(context_input: ContextGeneratorInput) -> Run[ContextGeneratorOutput]:
+async def generate_chunk_context(context_input: ContextGeneratorInput) -> ContextGeneratorOutput:
     """
     Here is the chunk we want to situate within the whole document.
     Please give a short succinct context to situate this chunk within the overall document
@@ -80,9 +80,9 @@ async def main():
         chunk_content=chunk_content,
     )
 
-    run = await generate_chunk_context(context_input)
+    run = await generate_chunk_context.run(context_input)
     print("\nGenerated Context:")
-    print(run.output.context)
+    print(run)
 
 
 if __name__ == "__main__":
diff --git a/examples/14_templated_instructions.py b/examples/14_templated_instructions.py
index 650f2b4..5b8bc31 100644
--- a/examples/14_templated_instructions.py
+++ b/examples/14_templated_instructions.py
@@ -23,7 +23,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class CodeReviewInput(BaseModel):
@@ -77,7 +77,7 @@ class CodeReviewOutput(BaseModel):
     id="templated-code-reviewer",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def review_code(review_input: CodeReviewInput) -> Run[CodeReviewOutput]:
+async def review_code(review_input: CodeReviewInput) -> CodeReviewOutput:
     """
     Review code based on specified parameters and guidelines.
 
@@ -142,7 +142,7 @@ def calculate_sum(numbers):
     return result
     """
 
-    run = await review_code(
+    run = await review_code.run(
         CodeReviewInput(
             language="python",
             code=python_code,
diff --git a/examples/15_text_to_sql.py b/examples/15_text_to_sql.py
index 908b61e..8257f61 100644
--- a/examples/15_text_to_sql.py
+++ b/examples/15_text_to_sql.py
@@ -17,7 +17,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class SQLGenerationInput(BaseModel):
@@ -49,7 +49,7 @@ class SQLGenerationOutput(BaseModel):
     id="text-to-sql",
     model=Model.CLAUDE_3_5_SONNET_LATEST,
 )
-async def generate_sql(review_input: SQLGenerationInput) -> Run[SQLGenerationOutput]:
+async def generate_sql(review_input: SQLGenerationInput) -> SQLGenerationOutput:
     """
     Convert natural language questions to SQL queries based on the provided schema.
 
@@ -121,7 +121,7 @@ async def main():
     # Example 1: Simple SELECT with conditions
     print("\nExample 1: Find expensive products")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question="Show me all products that cost more than $100, ordered by price descending",
@@ -132,7 +132,7 @@ async def main():
     # Example 2: JOIN with aggregation
     print("\nExample 2: Customer order summary")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question=(
@@ -146,7 +146,7 @@ async def main():
     # Example 3: Complex query
     print("\nExample 3: Product category analysis")
     print("-" * 50)
-    run = await generate_sql(
+    run = await generate_sql.run(
         SQLGenerationInput(
             db_schema=schema,
             question=(
diff --git a/examples/16_multi_model_consensus.py b/examples/16_multi_model_consensus.py
index da0fbbf..ba66f1c 100644
--- a/examples/16_multi_model_consensus.py
+++ b/examples/16_multi_model_consensus.py
@@ -15,7 +15,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class MultiModelInput(BaseModel):
@@ -56,7 +56,7 @@ class CombinedOutput(BaseModel):
 @workflowai.agent(
     id="question-answerer",
 )
-async def get_model_response(query: MultiModelInput) -> Run[ModelResponse]:
+async def get_model_response(query: MultiModelInput) -> ModelResponse:
     """Get response from the specified model."""
     ...
 
@@ -65,7 +65,7 @@ async def get_model_response(query: MultiModelInput) -> Run[ModelResponse]:
     id="response-combiner",
     model=Model.O3_MINI_2025_01_31_MEDIUM_REASONING_EFFORT,
 )
-async def combine_responses(responses_input: CombinerInput) -> Run[CombinedOutput]:
+async def combine_responses(responses_input: CombinerInput) -> CombinedOutput:
     """
     Analyze and combine responses from multiple models into a single coherent answer.
 
@@ -100,7 +100,7 @@ async def main():
 
     responses = []
     for model, model_name in models:
-        run = await get_model_response(
+        run = await get_model_response.run(
             MultiModelInput(
                 question=question,
                 model_name=model_name,
@@ -110,7 +110,7 @@ async def main():
         responses.append(run.output)
 
     # Combine responses
-    combined = await combine_responses(CombinerInput(responses=responses))
+    combined = await combine_responses.run(CombinerInput(responses=responses))
     print(combined)
 
 
diff --git a/examples/17_multi_model_consensus_with_tools.py b/examples/17_multi_model_consensus_with_tools.py
index 761e284..82806f4 100644
--- a/examples/17_multi_model_consensus_with_tools.py
+++ b/examples/17_multi_model_consensus_with_tools.py
@@ -23,7 +23,7 @@
 from pydantic import BaseModel, Field
 
 import workflowai
-from workflowai import Model, Run
+from workflowai import Model
 
 
 class AskModelInput(BaseModel):
@@ -44,13 +44,14 @@ class AskModelOutput(BaseModel):
 # through the get_model_response agent. This creates a hierarchy where the
 # response-combiner orchestrates multiple model queries by delegating to get_model_response.
 async def ask_model(query_input: AskModelInput) -> AskModelOutput:
-    """Ask a specific model a question and get its response."""
-    run = await get_model_response(
+    """Ask a specific model a question and return its response."""
+    run = await get_model_response.run(
         MultiModelInput(
             question=query_input.question,
         ),
         model=query_input.model,
     )
+    # get_model_response.run() returns a Run[ModelResponse], so we need to access the output
     return AskModelOutput(response=run.output.response)
 
 
@@ -91,7 +92,7 @@ class CombinedOutput(BaseModel):
 @workflowai.agent(
     id="question-answerer",
 )
-async def get_model_response(query: MultiModelInput) -> Run[ModelResponse]:
+async def get_model_response(query: MultiModelInput) -> ModelResponse:
     """
     Make sure to:
     1. Provide a clear and detailed response
@@ -107,7 +108,7 @@ async def get_model_response(query: MultiModelInput) -> Run[ModelResponse]:
     model=Model.GPT_4O_MINI_LATEST,
     tools=[ask_model],
 )
-async def combine_responses(responses_input: CombinerInput) -> Run[CombinedOutput]:
+async def combine_responses(responses_input: CombinerInput) -> CombinedOutput:
     """
     Analyze and combine responses from multiple models into a single coherent answer.
     You should ask at least 3 different models to get a diverse set of perspectives.
@@ -137,7 +138,7 @@ async def main():
     question = "What is dark matter and why is it important for our understanding of the universe?"
 
     # Let the response-combiner handle asking the models
-    combined = await combine_responses(
+    combined = await combine_responses.run(
         CombinerInput(
             original_question=question,
         ),