Error while doing research with Deep Research toggle on

[inderpreet001]

When using Deep Research feature, i am getting error every time the agent performs deep research. I have set the LiteLLm config.yml file as per the docs in ii-researcher, and also started the liteLLm server. I am uisng claude-sonnet model with openai and SerpApi key set in config.

[gauravdhiman]

+1 - I also got similar error when I tried running with deep research. Using Google Gemini as model.

Here are the backend server logs:

2025-08-20 23:21:59.619 | INFO:     Started server process [20]
2025-08-20 23:21:59.619 | INFO:     Waiting for application startup.
2025-08-20 23:21:59.619 | INFO:     Application startup complete.
2025-08-20 23:21:59.619 | INFO:     Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)
2025-08-20 23:22:03.603 | INFO:     ('172.64.149.246', 23515) - "WebSocket /ws?device_id=57f9b6e4-daf7-4138-9174-abf4c474e768" [accepted]
2025-08-20 23:22:03.604 | INFO  [ii_agent.server.websocket.manager] New WebSocket connection and chat session established: 281472296341504
2025-08-20 23:22:03.606 | INFO:     connection open
2025-08-20 23:22:03.608 | INFO:     172.64.149.246:17798 - "GET /api/sessions/57f9b6e4-daf7-4138-9174-abf4c474e768 HTTP/1.1" 200 OK
2025-08-20 23:22:03.610 | INFO:     172.64.149.246:64227 - "GET /api/settings HTTP/1.1" 200 OK
2025-08-20 23:22:58.202 | INFO:     172.64.149.246:28830 - "OPTIONS /api/settings HTTP/1.1" 200 OK
2025-08-20 23:22:58.213 | INFO:     172.64.149.246:28830 - "POST /api/settings HTTP/1.1" 200 OK
2025-08-20 23:28:45.784 | ====== Using Gemini directly ======
2025-08-20 23:28:45.792 | INFO  [ii_agent.server.websocket.chat_session] Created new session 083ef76d-6303-4ea5-a2fc-a130ca5cabfc with workspace at /.ii_agent/workspace/083ef76d-6303-4ea5-a2fc-a130ca5cabfc
2025-08-20 23:28:45.979 | Container created: fec19f18728fecd606a7981a9c024679a150240b0c87bfc2197294e72ebe1297
2025-08-20 23:28:45.980 | Using DuckDuckGo to search
2025-08-20 23:28:45.988 | /app/src/ii_agent/tools/clients/web_search_client.py:155: RuntimeWarning: This package (`duckduckgo_search`) has been renamed to `ddgs`! Use `pip install ddgs` instead.
2025-08-20 23:28:45.988 |   self.ddgs = DDGS(**kwargs)
2025-08-20 23:28:45.991 | Using Markdownify to visit webpage
2025-08-20 23:28:45.991 | No image search API key found
2025-08-20 23:28:45.992 | No history found for session 083ef76d-6303-4ea5-a2fc-a130ca5cabfc
2025-08-20 23:28:45.994 | /app/src/ii_agent/tools/clients/web_search_client.py:155: RuntimeWarning: This package (`duckduckgo_search`) has been renamed to `ddgs`! Use `pip install ddgs` instead.
2025-08-20 23:28:45.994 |   self.ddgs = DDGS(**kwargs)
2025-08-20 23:28:45.994 | Using DuckDuckGo to search
2025-08-20 23:28:45.994 | Using Markdownify to visit webpage
2025-08-20 23:28:45.994 | No image search API key found
2025-08-20 23:28:45.995 | Initialized Reviewer
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | --------------------------------------------- USER INPUT ---------------------------------------------
2025-08-20 23:28:46.004 | I'm looking for very detailed and thorough research about long-running, multi-turn agent loops using AI, especially in the open-source ecosystem.
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | There are some good systems from China, like Minimax Agent, Manus AI, and Genspark. What are some open-source alternatives to those that are equally good or better, robust, and LLM-agnostic, which can actually plan and execute the plan?
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | These systems normally have access to a terminal, a browser, or a complete computer running in a container or virtual environment. And their systems are very generic but highly capable in software engineering, deep market research, presentation preparation, etc.
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | I'm mainly interested in the open-source solutions, but the ones that are robust and are active projects with high potential as an alternative to managed AI like GenSpark or Minimax agents.
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | Generate a beautiful report after your research, with HTML, JS and CSS and host it.
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.004 | --------------------------------------------- NEW TURN ---------------------------------------------
2025-08-20 23:28:46.004 | 
2025-08-20 23:28:46.006 | (Current token count: 308)
2025-08-20 23:28:46.006 | 
2025-08-20 23:28:58.723 | WARNI [google_genai.types] Warning: there are non-text parts in the response: ['thought_signature', 'function_call'], returning concatenated text result from text parts. Check the full candidates.content.parts accessor to get the full model response.
2025-08-20 23:28:58.724 | Running tool: message_user
2025-08-20 23:28:58.724 | Tool input: {'text': 'I will start by conducting in-depth research on long-running, multi-turn agent loops in the open-source ecosystem. I will then identify and analyze robust, LLM-agnostic alternatives to closed-source systems like Minimax Agent, Manus AI, and Genspark. Following the research, I will create a beautiful and detailed report with HTML, JS, and CSS, host it, and provide you with the link.'}
2025-08-20 23:28:58.725 | Calling tool message_user with input:
2025-08-20 23:28:58.725 |  - text: I will start by conducting in-depth research on long-running, multi-turn agent loops in the open-source ecosystem. I will then identify and analyze robust, LLM-agnostic alternatives to closed-source systems like Minimax Agent, Manus AI, and Genspark. Following the research, I will create a beautiful and detailed report with HTML, JS, and CSS, host it, and provide you with the link.
2025-08-20 23:28:58.725 | Tool output: 
2025-08-20 23:28:58.725 | Sent message to user
2025-08-20 23:28:58.725 | 
2025-08-20 23:28:58.725 | 
2025-08-20 23:28:58.725 | 
2025-08-20 23:28:58.725 | --------------------------------------------- NEW TURN ---------------------------------------------
2025-08-20 23:28:58.725 | 
2025-08-20 23:28:58.725 | (Current token count: 446)
2025-08-20 23:28:58.725 | 
2025-08-20 23:29:11.725 | WARNI [google_genai.types] Warning: there are non-text parts in the response: ['thought_signature', 'function_call'], returning concatenated text result from text parts. Check the full candidates.content.parts accessor to get the full model response.
2025-08-20 23:29:11.726 | Running tool: deep_research
2025-08-20 23:29:11.726 | Tool input: {'query': 'long-running, multi-turn agent loops using AI in the open-source ecosystem, focusing on robust, LLM-agnostic alternatives to closed-source systems like Minimax Agent, Manus AI, and Genspark that can plan and execute tasks with access to terminals and browsers.'}
2025-08-20 23:29:11.727 | Performing deep research on long-running, multi-turn agent loops using AI in the open-source ecosystem, focusing on robust, LLM-agnostic alternatives to closed-source systems like Minimax Agent, Manus AI, and Genspark that can plan and execute tasks with access to terminals and browsers.
2025-08-20 23:29:13.086 | ERROR [root] Error generating streaming completion: Connection error.
2025-08-20 23:29:13.086 | ERROR [ii_agent.server.websocket.chat_session] Error running agent: Connection error.
2025-08-20 23:29:13.093 | Traceback (most recent call last):
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_transports/default.py", line 101, in map_httpcore_exceptions
2025-08-20 23:29:13.093 |     yield
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_transports/default.py", line 394, in handle_async_request
2025-08-20 23:29:13.093 |     resp = await self._pool.handle_async_request(req)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_async/connection_pool.py", line 256, in handle_async_request
2025-08-20 23:29:13.093 |     raise exc from None
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_async/connection_pool.py", line 236, in handle_async_request
2025-08-20 23:29:13.093 |     response = await connection.handle_async_request(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_async/connection.py", line 101, in handle_async_request
2025-08-20 23:29:13.093 |     raise exc
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_async/connection.py", line 78, in handle_async_request
2025-08-20 23:29:13.093 |     stream = await self._connect(request)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_async/connection.py", line 124, in _connect
2025-08-20 23:29:13.093 |     stream = await self._network_backend.connect_tcp(**kwargs)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_backends/auto.py", line 31, in connect_tcp
2025-08-20 23:29:13.093 |     return await self._backend.connect_tcp(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_backends/anyio.py", line 113, in connect_tcp
2025-08-20 23:29:13.093 |     with map_exceptions(exc_map):
2025-08-20 23:29:13.093 |   File "/root/.local/share/uv/python/cpython-3.10.18-linux-aarch64-gnu/lib/python3.10/contextlib.py", line 153, in __exit__
2025-08-20 23:29:13.093 |     self.gen.throw(typ, value, traceback)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpcore/_exceptions.py", line 14, in map_exceptions
2025-08-20 23:29:13.093 |     raise to_exc(exc) from exc
2025-08-20 23:29:13.093 | httpcore.ConnectError: All connection attempts failed
2025-08-20 23:29:13.093 | 
2025-08-20 23:29:13.093 | The above exception was the direct cause of the following exception:
2025-08-20 23:29:13.093 | 
2025-08-20 23:29:13.093 | Traceback (most recent call last):
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/openai/_base_client.py", line 1519, in request
2025-08-20 23:29:13.093 |     response = await self._client.send(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_client.py", line 1629, in send
2025-08-20 23:29:13.093 |     response = await self._send_handling_auth(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_client.py", line 1657, in _send_handling_auth
2025-08-20 23:29:13.093 |     response = await self._send_handling_redirects(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_client.py", line 1694, in _send_handling_redirects
2025-08-20 23:29:13.093 |     response = await self._send_single_request(request)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_client.py", line 1730, in _send_single_request
2025-08-20 23:29:13.093 |     response = await transport.handle_async_request(request)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_transports/default.py", line 393, in handle_async_request
2025-08-20 23:29:13.093 |     with map_httpcore_exceptions():
2025-08-20 23:29:13.093 |   File "/root/.local/share/uv/python/cpython-3.10.18-linux-aarch64-gnu/lib/python3.10/contextlib.py", line 153, in __exit__
2025-08-20 23:29:13.093 |     self.gen.throw(typ, value, traceback)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/httpx/_transports/default.py", line 118, in map_httpcore_exceptions
2025-08-20 23:29:13.093 |     raise mapped_exc(message) from exc
2025-08-20 23:29:13.093 | httpx.ConnectError: All connection attempts failed
2025-08-20 23:29:13.093 | 
2025-08-20 23:29:13.093 | The above exception was the direct cause of the following exception:
2025-08-20 23:29:13.093 | 
2025-08-20 23:29:13.093 | Traceback (most recent call last):
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/server/websocket/chat_session.py", line 741, in _run_agent_async
2025-08-20 23:29:13.093 |     await self.agent.run_agent_async(user_input, files, resume)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/agents/function_call.py", line 367, in run_agent_async
2025-08-20 23:29:13.093 |     return await self.run_async(tool_input, self.history)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/tools/base.py", line 69, in run_async
2025-08-20 23:29:13.093 |     result = await self.run_impl(tool_input, message_history)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/agents/function_call.py", line 314, in run_impl
2025-08-20 23:29:13.093 |     tool_result = await self.tool_manager.run_tool(tool_call, self.history)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/tools/tool_manager.py", line 363, in run_tool
2025-08-20 23:29:13.093 |     result = await llm_tool.run_async(tool_input, history)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/tools/base.py", line 69, in run_async
2025-08-20 23:29:13.093 |     result = await self.run_impl(tool_input, message_history)
2025-08-20 23:29:13.093 |   File "/app/src/ii_agent/tools/deep_research_tool.py", line 63, in run_impl
2025-08-20 23:29:13.093 |     result = await agent.run(on_token=on_token, is_stream=True)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/ii_researcher/reasoning/agent.py", line 113, in run
2025-08-20 23:29:13.093 |     content = await self.process_stream(stream_generator, on_token)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/ii_researcher/reasoning/agent.py", line 63, in process_stream
2025-08-20 23:29:13.093 |     async for token in stream_generator:
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/ii_researcher/reasoning/clients/openai_client.py", line 81, in generate_completion_stream
2025-08-20 23:29:13.093 |     stream = await self.async_client.chat.completions.create(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/openai/resources/chat/completions/completions.py", line 2454, in create
2025-08-20 23:29:13.093 |     return await self._post(
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/openai/_base_client.py", line 1784, in post
2025-08-20 23:29:13.093 |     return await self.request(cast_to, opts, stream=stream, stream_cls=stream_cls)
2025-08-20 23:29:13.093 |   File "/app/.venv/lib/python3.10/site-packages/openai/_base_client.py", line 1551, in request
2025-08-20 23:29:13.093 |     raise APIConnectionError(request=request) from err
2025-08-20 23:29:13.093 | openai.APIConnectionError: Connection error.

[bruceye777]

+1 same error
backend-1 | File "/app/.venv/lib/python3.10/site-packages/openai/_base_client.py", line 1561, in request
backend-1 | raise APIConnectionError(request=request) from err
backend-1 | openai.APIConnectionError: Connection error.