[BUG] TypeError: Object of type File is not JSON serializable when using input_files with gemma3 #4498
Description
Description
When using the crewai_files module to pass File objects into crew.kickoff(input_files=...), the execution fails during the LLM call. The underlying httpx client attempts to serialize the File object into a JSON payload for the OpenAI-compatible endpoint (Ollama/Gemma 3), leading to a TypeError.
Steps to Reproduce
1- Define a Crew with an agent using a local LLM (e.g., Gemma 3).
2- Create a File object using crewai_files.
3- Call crew.kickoff(input_files={"my_file": File(source="path/to/file.pdf")}).
4- See error: TypeError: Object of type File is not JSON serializable.
Expected behavior
The File object should be correctly handled by the agent's executor.
Screenshots/Code snippets
ERROR: OpenAI API call failed: Object of type File is not JSON serializable
ERROR: OpenAI API call failed: Object of type File is not JSON serializable
[CrewAIEventsBus] Warning: Event pairing mismatch. 'llm_call_failed' closed 'agent_execution_started' (expected 'llm_call_started')
An unknown error occurred. Please check the details below.
Error details: Object of type File is not JSON serializable
An unknown error occurred. Please check the details below.
Error details: Object of type File is not JSON serializable
[CrewAIEventsBus] Warning: Event pairing mismatch. 'agent_execution_error' closed 'task_started' (expected
'agent_execution_started')
[CrewAIEventsBus] Warning: Event pairing mismatch. 'task_failed' closed 'crew_kickoff_started' (expected 'task_started')
[CrewAIEventsBus] Warning: Ending event 'crew_kickoff_failed' emitted with empty scope stack. Missing starting event?
ERROR: Programmatic analysis failed for 138/2026
Traceback (most recent call last):
File "/.../project/src/ai_albo_bot/telegram/notifier.py", line 207, in notify_users
res = analyze.analyze_item(entry)
File "/.../project/src/ai_albo_bot/agent/analyze.py", line 137, in analyze_item
crew_result = crew.kickoff(input_files=input_files)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/crew.py", line 743, in kickoff
result = self._run_sequential_process()
File "/.../project/.venv/lib/python3.13/site-packages/crewai/crew.py", line 1150, in _run_sequential_process
return self._execute_tasks(self.tasks)
~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/crew.py", line 1236, in _execute_tasks
task_output = task.execute_sync(
agent=exec_data.agent,
context=context,
tools=exec_data.tools,
)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/task.py", line 499, in execute_sync
return self._execute_core(agent, context, tools)
~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/task.py", line 740, in _execute_core
raise e # Re-raise the exception after emitting the event
^^^^^^^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/task.py", line 671, in _execute_core
result = agent.execute_task(
task=self,
context=context,
tools=tools,
)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agent/core.py", line 493, in execute_task
result = self.execute_task(task, context, tools)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agent/core.py", line 493, in execute_task
result = self.execute_task(task, context, tools)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agent/core.py", line 492, in execute_task
raise e
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agent/core.py", line 459, in execute_task
result = self._execute_without_timeout(task_prompt, task)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agent/core.py", line 568, in _execute_without_timeout
return self.agent_executor.invoke(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
{
^
...<4 lines>...
}
^
)["output"]
^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agents/crew_agent_executor.py", line 207, in invoke
formatted_answer = self._invoke_loop()
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agents/crew_agent_executor.py", line 305, in _invoke_loop
return self._invoke_loop_native_tools()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agents/crew_agent_executor.py", line 579, in _invoke_loop_native_tools
raise e
File "/.../project/.venv/lib/python3.13/site-packages/crewai/agents/crew_agent_executor.py", line 498, in _invoke_loop_native_tools
answer = get_llm_response(
llm=self.llm,
...<9 lines>...
verbose=self.agent.verbose,
)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/utilities/agent_utils.py", line 372, in get_llm_response
raise e
File "/.../project/.venv/lib/python3.13/site-packages/crewai/utilities/agent_utils.py", line 362, in get_llm_response
answer = llm.call(
messages,
...<5 lines>...
response_model=response_model,
)
File "/.../project/.venv/lib/python3.13/site-packages/crewai/llms/providers/openai/completion.py", line 410, in call
return self._call_completions(
~~~~~~~~~~~~~~~~~~~~~~^
messages=formatted_messages,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...<4 lines>...
response_model=response_model,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/llms/providers/openai/completion.py", line 450, in _call_completions
return self._handle_completion(
~~~~~~~~~~~~~~~~~~~~~~~^
params=completion_params,
^^^^^^^^^^^^^^^^^^^^^^^^^
...<3 lines>...
response_model=response_model,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/.../project/.venv/lib/python3.13/site-packages/crewai/llms/providers/openai/completion.py", line 1687, in _handle_completion
raise e from e
File "/.../project/.venv/lib/python3.13/site-packages/crewai/llms/providers/openai/completion.py", line 1585, in _handle_completion
response: ChatCompletion = self.client.chat.completions.create(**params)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^
File "/.../project/.venv/lib/python3.13/site-packages/openai/_utils/_utils.py", line 287, in wrapper
return func(*args, **kwargs)
File "/.../project/.venv/lib/python3.13/site-packages/openai/resources/chat/completions/completions.py", line 925, in create
return self._post(
~~~~~~~~~~^
"/chat/completions",
^^^^^^^^^^^^^^^^^^^^
...<43 lines>...
stream_cls=Stream[ChatCompletionChunk],
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/.../project/.venv/lib/python3.13/site-packages/openai/_base_client.py", line 1242, in post
return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/.../project/.venv/lib/python3.13/site-packages/openai/_base_client.py", line 958, in request
request = self._build_request(options, retries_taken=retries_taken)
File "/.../project/.venv/lib/python3.13/site-packages/openai/_base_client.py", line 535, in _build_request
return self._client.build_request( # pyright: ignore[reportUnknownMemberType]
~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
headers=headers,
^^^^^^^^^^^^^^^^
...<10 lines>...
**kwargs,
^^^^^^^^^
)
^
File "/.../project/.venv/lib/python3.13/site-packages/httpx/_client.py", line 378, in build_request
return Request(
method,
...<8 lines>...
extensions=extensions,
)
File "/.../project/.venv/lib/python3.13/site-packages/httpx/_models.py", line 408, in __init__
headers, stream = encode_request(
~~~~~~~~~~~~~~^
content=content,
^^^^^^^^^^^^^^^^
...<7 lines>...
),
^^
)
^
File "/.../project/.venv/lib/python3.13/site-packages/httpx/_content.py", line 216, in encode_request
return encode_json(json)
File "/.../project/.venv/lib/python3.13/site-packages/httpx/_content.py", line 177, in encode_json
body = json_dumps(
~~~~~~~~~~^
json, ensure_ascii=False, separators=(",", ":"), allow_nan=False
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
).encode("utf-8")
^
File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/json/__init__.py", line 242, in dumps
**kw).encode(obj)
~~~~~~^^^^^
File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/json/encoder.py", line 202, in encode
chunks = self.iterencode(o, _one_shot=True)
File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/json/encoder.py", line 263, in iterencode
return _iterencode(o, 0)
File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/json/encoder.py", line 182, in default
raise TypeError(f'Object of type {o.__class__.__name__} '
f'is not JSON serializable')
TypeError: Object of type File is not JSON serializable
Operating System
macOS 26
Python Version
3.13
crewAI Version
1.9.3
crewAI Tools Version
1.9.3
Virtual Environment
Venv
Evidence
INFO: OpenAI: Successfully validated tool 'read_file'
ERROR: OpenAI API call failed: Object of type File is not JSON serializable
ERROR: OpenAI API call failed: Object of type File is not JSON serializable
[CrewAIEventsBus] Warning: Event pairing mismatch. 'llm_call_failed' closed 'agent_execution_started' (expected 'llm_call_started')
An unknown error occurred. Please check the details below.
Error details: Object of type File is not JSON serializable
An unknown error occurred. Please check the details below.
Error details: Object of type File is not JSON serializable
Possible Solution
None
Additional context
LLM: gemma3 via Docker Model Runner