[GH-ISSUE #5670] [BUG]: Agent web search errors out on clear chat #5229

Closed
opened 2026-06-05 14:52:48 -04:00 by yindo · 2 comments
Owner

Originally created by @Hallavera on GitHub (May 21, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5670

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

When starting a new chat I immediately call agent to do a web search for me.
At this point it throws a "failed to parse error".

If I write any message in the chat first then the agent is succesful.

Are there known steps to reproduce?

Image Image

LLM Provider & Model (if applicable)

gpt-oss-120b
qwen3.5-35B-A3B

Embedder Provider & Model (if applicable)

No response

Originally created by @Hallavera on GitHub (May 21, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5670 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? When starting a new chat I immediately call agent to do a web search for me. At this point it throws a "failed to parse error". If I write any message in the chat first then the agent is succesful. ### Are there known steps to reproduce? <img width="1896" height="532" alt="Image" src="https://github.com/user-attachments/assets/8f580e37-5ca2-4b84-ac9e-0c1f20e84089" /> <img width="1898" height="722" alt="Image" src="https://github.com/user-attachments/assets/229c34f2-e366-4a38-91fe-d9aa7d9f9962" /> ### LLM Provider & Model (if applicable) gpt-oss-120b qwen3.5-35B-A3B ### Embedder Provider & Model (if applicable) _No response_
yindo added the possible bug label 2026-06-05 14:52:48 -04:00
yindo closed this issue 2026-06-05 14:52:48 -04:00
Author
Owner

@timothycarambat commented on GitHub (May 21, 2026):

This is a bug/issue with your LLM Provider (not the model) as it is having an issue with the chat template/tokenizer causing it to crash during response streaming.

This is not something we can fix on our end - what are you using for the provider and do you have to manually manage the chat template or anything like this? That is what this error looks like not know much else

<!-- gh-comment-id:4508286287 --> @timothycarambat commented on GitHub (May 21, 2026): This is a bug/issue with your LLM **Provider** (not the model) as it is having an issue with the chat template/tokenizer causing it to crash during response streaming. This is not something we can fix on our end - what are you using for the provider and do you have to manually manage the chat template or anything like this? That is what this error looks like not know much else
Author
Owner

@Hallavera commented on GitHub (May 21, 2026):

Im using llama-cpp and specifying only host and port.

https://github.com/ggml-org/llama.cpp/issues/20650#issue-4084325944

Seems to be this issue. Sorry I was not able to find it before.

<!-- gh-comment-id:4508803829 --> @Hallavera commented on GitHub (May 21, 2026): Im using llama-cpp and specifying only host and port. https://github.com/ggml-org/llama.cpp/issues/20650#issue-4084325944 Seems to be this issue. Sorry I was not able to find it before.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#5229