(Un)Perplexed Spready is a spreadsheet software that thinks! It combines familiar spreadsheet power with revolutionary AI models for effortless data analysis and smart insights. Use AI-powered formulas to unlock insights beyond the reach of Excel. Unleash Artificial Intelligence in your spreadsheets! Drink you coffee, while the AI is doing hard job for you! You can now try it online for free!

(Un)Perplexed Spready is a spreadsheet software that thinks! It combines familiar spreadsheet power with revolutionary AI models for effortless data analysis and smart insights. Use AI-powered formulas to unlock insights beyond the reach of Excel. Unleash Artificial Intelligence in your spreadsheets! Drink you coffee, while the AI is doing hard job for you! You can now try it online for free!

(Un)Perplexed Spready is a spreadsheet software that thinks! It combines familiar spreadsheet power with revolutionary AI models for effortless data analysis and smart insights. Use AI-powered formulas to unlock insights beyond the reach of Excel. Unleash Artificial Intelligence in your spreadsheets! Drink you coffee, while the AI is doing hard job for you! You can now try it online for free!

Hi the problem is resolved, thanks for your response, it seems that the SSL or TLS handshake doesn’t work properly, and I tried to go to the duckduckgo website and it returns error. But now it is solved, the problem maybe lies in the date and time of the system which is still not in sync with my local time (as I am currently in a different time zone). The other approach is maybe to clear the SSL state

The content seems strange, or rather, it looks like the query isn’t being passed… There are several implementations of search tools, but if it’s only happening with one of them, the search engine specifications may have changed and the library isn’t compatible. github.com/huggingface/smolagents #### WebSearchTool example from Guide Tour does not work opened 08:56PM - 26 May 25 UTC closed 07:27AM - 27 May 25 UTC AlexiaJM bug **Describe the bug** The example about web search from the Guided Tour does not …work. I have internet access. **Code to reproduce the error** > from smolagents import WebSearchTool > search_tool = WebSearchTool() > print(search_tool("Who is the president of Russia?")) **Error logs (if any)** > Traceback (most recent call last): File "<stdin>", line 1, in <module> File "env_home/lib/python3.10/site-packages/smolagents/tools.py", line 205, in __call__ outputs = self.forward(*args, **kwargs) File "env_home/lib/python3.10/site-packages/smolagents/default_tools.py", line 227, in forward raise Exception("No results found! Try a less restrictive/shorter query.") Exception: No results found! Try a less restrictive/shorter query. **Packages version:** smolagents==1.16.1

The content seems strange, or rather, it looks like the query isn’t being passed… There are several implementations of search tools, but if it’s only happening with one of them, the search engine specifications may have changed and the library isn’t compatible. github.com/huggingface/smolagents #### WebSearchTool example from Guide Tour does not work opened 08:56PM - 26 May 25 UTC closed 07:27AM - 27 May 25 UTC AlexiaJM bug **Describe the bug** The example about web search from the Guided Tour does not …work. I have internet access. **Code to reproduce the error** > from smolagents import WebSearchTool > search_tool = WebSearchTool() > print(search_tool("Who is the president of Russia?")) **Error logs (if any)** > Traceback (most recent call last): File "<stdin>", line 1, in <module> File "env_home/lib/python3.10/site-packages/smolagents/tools.py", line 205, in __call__ outputs = self.forward(*args, **kwargs) File "env_home/lib/python3.10/site-packages/smolagents/default_tools.py", line 227, in forward raise Exception("No results found! Try a less restrictive/shorter query.") Exception: No results found! Try a less restrictive/shorter query. **Packages version:** smolagents==1.16.1

The content seems strange, or rather, it looks like the query isn’t being passed… There are several implementations of search tools, but if it’s only happening with one of them, the search engine specifications may have changed and the library isn’t compatible. github.com/huggingface/smolagents #### WebSearchTool example from Guide Tour does not work opened 08:56PM - 26 May 25 UTC closed 07:27AM - 27 May 25 UTC AlexiaJM bug **Describe the bug** The example about web search from the Guided Tour does not …work. I have internet access. **Code to reproduce the error** > from smolagents import WebSearchTool > search_tool = WebSearchTool() > print(search_tool("Who is the president of Russia?")) **Error logs (if any)** > Traceback (most recent call last): File "<stdin>", line 1, in <module> File "env_home/lib/python3.10/site-packages/smolagents/tools.py", line 205, in __call__ outputs = self.forward(*args, **kwargs) File "env_home/lib/python3.10/site-packages/smolagents/default_tools.py", line 227, in forward raise Exception("No results found! Try a less restrictive/shorter query.") Exception: No results found! Try a less restrictive/shorter query. **Packages version:** smolagents==1.16.1

I tried it, it is working now haha, at least it can surf the internet, but the result still need to be finetuned i think, thanks for the recommendation

I tried it, it is working now haha, at least it can surf the internet, but the result still need to be finetuned i think, thanks for the recommendation

I tried the smolagents WebSearchTool to search some information, but it returns irrelevant information, I don’t know if there is a way to fine-tune the result or the query, attached is the code generated from smolagents and the result

Hmm… For example, how about with `WebSearchTool(engine="bing")` ? github.com/huggingface/smolagents #### src/smolagents/default_tools.py `v1.19.0` 249. if not results: 250. return "No results found." 251. return "## Search Results\n\n" + "\n\n".join( 252. [ 253. f"{idx}. {result['title']}\n{result['description']}" 254. for idx, result in enumerate(results, start=1) 255. ] 256. ) 257. 258. 259. class WebSearchTool(Tool): 260. name = "web_search" 261. description = "Performs a web search for a query and returns a string of the top search results formatted as markdown with titles, links, and descriptions." 262. inputs = {"query": {"type": "string", "description": "The search query to perform."}} 263. output_type = "string" 264. 265. def __init__(self, max_results: int = 10, engine: str = "duckduckgo"): 266. super().__init__() 267. self.max_results = max_results 268. self.engine = engine 269.

have you figured out the solution yet, cause I solved the SSL issue already but stuck with the same problem as you

Hi, thanks for answering, I tried the StackOverflow solution already, the issue seems to be solved, but now I got max retries exceeded error, I still try to find the solution for it

Hmm… Perhaps DDG problem…? GitHub ### Can't Get Web Search DuckDuckGo Working · open-webui open-webui · Discussion... Bug Report Installation Method pip install openwebui ollama Environment Open WebUI Version: [e.g., v0.3.11] Ollama (if applicable): [e.g., v0.2.0, v0.1.32-rc1] Operating System: [e.g., Windows 10, ...

I’m having a similar error with DuckDuckGo Code execution failed at line ‘results_retry = web_search(query=simpler_query)’ due to: DuckDuckGoSearchException: DuckDuckGo RuntimeError: error sending request for url (DuckDuckGo): operation timed out Caused by: operation timed out

I think this might be an SSL error caused by a proxy, VPN, cloud, or internal network firewall, but it’s in the library code… It might be difficult to work around. huggingface.co ### Tools We’re on a journey to advance and democratize artificial intelligence through open source and open science. https://stackoverflow.com/questions/51925384/unable-to-get-local-issuer-certificate-when-using-requests

Hi I tried to use WebSearchTool from smolagents and got this kind of error, I’m using ollama with model qwen2.5 7b, can anyone help me Code execution failed at line ‘music_recommendations = web_search(query=“best party music”)’ due to: SSLError: HTTPSConnectionPool(host=‘lite.duckduckgo.com’, port=443): Max retries exceeded with url: /lite/?q=best+party+music (Caused by SSLError(SSLCertVerificationError(1, ‘[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate (_ssl.c:1028)’)))

I think you probably forgot to run `pip install` in the first cell, causing the `huggingface_hub` library to be outdated, and also because of a specification change in `smolagents`. It’s probably a library version issue. from smolagents import CodeAgent, WebSearchTool, InferenceClientModel agent = CodeAgent(tools=[WebSearchTool()], model=InferenceClientModel()) agent.run("Search for the best music recommendations for a party at the Wayne's mansion.") github.com/huggingface/agents-course #### [HANDS-ON BUG] DuckDuckGoSearchTool use in smolagents notebook fails opened 09:16PM - 12 Jun 25 UTC timadair hands-on-bug **Describe the bug** When running the code in Building Agents That Use Codetps://huggingface.co/learn/agents-course/unit2/smolagents/code_agents), the DuckDuckGo search tool is no longer included in the default smolagents package. **To Reproduce** Run the code in the notebook. The failure happens in the third code block. **Additional context** It appears that the DuckDuckGoSearchTool was removed as a required dependency in smolagents v1.15.0, and was replaced in the documentation with WebSearchTool. Remember to update the course pages and not just the notebook. github.com/huggingface/huggingface_hub #### ValueError: Provider 'featherless-ai' not supported. opened 09:58AM - 09 May 25 UTC closed 03:45PM - 12 May 25 UTC 0x2b3bfa0 bug ### Describe the bug It looks like `GET /api/models/...?expand=inferenceProvide…rMapping` is returning a still unsupported inference provider. https://github.com/huggingface/huggingface_hub/blob/785835fc58defeff40368663589cca0a9a8808da/src/huggingface_hub/inference/_providers/__init__.py#L52-L131 Would it be more convenient to skip unsupported providers instead of raising an exception? ### Reproduction #### `requirements.txt` ```console huggingface_hub==0.31.1 ``` #### `example.py` ```python from huggingface_hub import InferenceClient InferenceClient("meta-llama/Llama-3.1-70B-Instruct").chat_completion(messages=[]) ``` ### Logs #### `python example.py` ```python Traceback (most recent call last): File "/.../example.py", line 13, in <module> InferenceClient("meta-llama/Llama-3.1-70B-Instruct").chat_completion(messages=[]) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^ File "/.../.venv/lib/python3.13/site-packages/huggingface_hub/inference/_client.py", line 886, in chat_completion provider_helper = get_provider_helper( self.provider, ...<3 lines>... else payload_model, ) File "/.../.venv/lib/python3.13/site-packages/huggingface_hub/inference/_providers/__init__.py", line 169, in get_provider_helper raise ValueError( ...<3 lines>... ) ValueError: Provider 'featherless-ai' not supported. Available values: 'auto' or any provider from ['black-forest-labs', 'cerebras', 'cohere', 'fal-ai', 'fireworks-ai', 'hf-inference', 'hyperbolic', 'nebius', 'novita', 'openai', 'replicate', 'sambanova', 'together'].Passing 'auto' (default value) will automatically select the first provider available for the model, sorted by the user's order in https://hf.co/settings/inference-providers. ``` #### <code>GET <a href="https://huggingface.co/api/models/meta-llama/Llama-3.1-70B-Instruct?expand=inferenceProviderMapping">https://huggingface.co/api/models/meta-llama/Llama-3.1-70B-Instruct?expand=inferenceProviderMapping</a></code> ```json { "_id": "66969ad27a033bf62173f3e2", "id": "meta-llama/Llama-3.1-70B-Instruct", "inferenceProviderMapping": { "featherless-ai": { "status": "live", "providerId": "meta-llama/Meta-Llama-3.1-70B-Instruct", "task": "conversational" }, "novita": { "status": "live", "providerId": "meta-llama/llama-3.1-70b-instruct", "task": "conversational" }, "nebius": { "status": "live", "providerId": "meta-llama/Meta-Llama-3.1-70B-Instruct-fast", "task": "conversational" }, "hyperbolic": { "status": "staging", "providerId": "meta-llama/Meta-Llama-3.1-70B-Instruct", "task": "conversational" } } } ``` ### System info ```console - huggingface_hub version: 0.31.0 - Platform: macOS-15.4.1-arm64-arm-64bit-Mach-O - Python version: 3.13.3 - Running in iPython ?: No - Running in notebook ?: No - Running in Google Colab ?: No - Running in Google Colab Enterprise ?: No - Token path ?: /.../.cache/huggingface/token - Has saved token ?: False - Configured git credential helpers: osxkeychain - FastAI: N/A - Tensorflow: N/A - Torch: 2.7.0 - Jinja2: 3.1.6 - Graphviz: N/A - keras: N/A - Pydot: N/A - Pillow: 11.2.1 - hf_transfer: 0.1.9 - gradio: N/A - tensorboard: N/A - numpy: 2.2.5 - pydantic: 2.10.6 - aiohttp: 3.11.18 - hf_xet: 1.1.0 - ENDPOINT: https://huggingface.co - HF_HUB_CACHE: /.../.cache/huggingface/hub - HF_ASSETS_CACHE: /.../.cache/huggingface/assets - HF_TOKEN_PATH: /.../.cache/huggingface/token - HF_STORED_TOKENS_PATH: /.../.cache/huggingface/stored_tokens - HF_HUB_OFFLINE: False - HF_HUB_DISABLE_TELEMETRY: False - HF_HUB_DISABLE_PROGRESS_BARS: None - HF_HUB_DISABLE_SYMLINKS_WARNING: False - HF_HUB_DISABLE_EXPERIMENTAL_WARNING: False - HF_HUB_DISABLE_IMPLICIT_TOKEN: False - HF_HUB_ENABLE_HF_TRANSFER: False - HF_HUB_ETAG_TIMEOUT: 10 - HF_HUB_DOWNLOAD_TIMEOUT: 10 ```

this issue still exists as of today 28-Dec-25. ** WebSearchTool with Structured Outputs causes JSON truncation (EOF error) - response grows unbounded** ### **Environment** * **Model** : gpt-4.1-mini-2025-04-14 * **API** : Responses API with WebSearchTool * **Output** : Structured Outputs (Pydantic schema) * **Parameters tested** : * max_output_tokens=32768 (model max) * search_context_size=“low” (minimum) ### **Problem Description** When using WebSearchTool with structured outputs, the API returns truncated JSON that fails Pydantic validation with: Invalid JSON: EOF while parsing a list at line X column Y The response is cut off mid-JSON, producing invalid output regardless of prompt instructions or parameter settings. ### **Reproduction Steps** 1. Create a Responses API request with: * WebSearchTool enabled * Structured output schema (Pydantic model) * A company/entity search query 1. Observe truncated JSON in response ### **Test Results** Attempt | Truncation Point | Duration | Prompt Strategy ---|---|---|--- 1 | column 6,791 | ~20s | Basic 2 | column 34,601 | ~3.8 min | Added max_output_tokens=32768 3 | column 48,433 | ~4.8 min | Added “keep response concise, max 3 items per field” **Key observation** : Adding instructions to be concise made the response **larger** , not smaller. The model appears to ignore output length guidance when web search fills the context. ### **Workarounds Attempted (None Fully Effective)** 1. Set max_output_tokens=32768 - Response grew but still truncated 2. Set search_context_size=“low” - No improvement 3. Added prompt: “Ensure JSON is valid and complete” - No improvement 4. Added prompt: “Maximum 3 items per field, prioritize quality” - Response grew larger 5. Simplified schema to fewer fields - Still truncates ### **Expected Behavior** * The API should return valid, parseable JSON that conforms to the schema * If context is exhausted, the model should gracefully close the JSON structure * search_context_size=“low” should limit internal context consumption ### **Actual Behavior** * Web search appears to consume unbounded context internally * Model generates partial JSON until some limit is hit * Response is truncated mid-structure with no closing brackets * No error returned - API reports success with invalid JSON payload ### **Workaround Implemented** We catch the Pydantic validation error and extract partial data from the truncated response for logging/debugging, but this is not a production-viable solution. ### **Related Issues** * Agent using WebSearchTool with structured outputs results in validation error with JSON unexpectedly ending with EOF around 6000 characters (March 2025) ### **Request** 1. Fix the internal context management for WebSearchTool 2. If truncation is unavoidable, ensure JSON is properly closed before returning 3. Return an appropriate error/warning when web search context exceeds limits

I have found that this is only an issue when I’m combining the WebSearchTool and enforcing a structured response with Pydantic. If you move to asking nicely for JSON (as opposed to forcing the output with Pydantic) these markers get replaced with proper markdown, just like I’m using to seeing with earlier models. That’s a workaround, but it’s not ideal since sometimes I get back malformed JSON objects that don’t match my spec. I’d love to see a fix for this from @OpenAI_Support ! Alternatively, clear documentation on how to parse the citations would be helpful.

I wish it were that simple. Luckily, OpenAI seems to have patched this bug. Json schemas now work with or without the extra instruction