mirror of
https://github.com/open-webui/computer.git
synced 2026-11-05 01:17:24 +00:00
[GH-ISSUE #125] feat: pass reasoning_content back to the model provider #62
Labels
No labels
enhancement
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
open-webui/computer#62
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @kierro-max on GitHub (Jul 10, 2026).
Original GitHub issue: https://github.com/open-webui/computer/issues/125
I am currently running cptr 0.8.8 and it is a great tool. I use Llama.cpp and Qwen 3.6 27B as the provider/model.
Ideally I would like to enable {preserve_thinking: true} on Llama.cpp to have Qwen remember past thought and make smarter decision.
It is not obvious to me if and how this can be accomplished with cptr. In Open WebUI it was a simple provider field in the connection window where I changed from default to Llama.cpp. The field does not exist in cptr.
@owui-terminator[bot] commented on GitHub (Jul 10, 2026):
⚠️ Missing Issue Title Prefix
@kierro-max, your issue title is missing a categorising prefix (e.g.
bug:,feat:,docs:).Please update the title to lead with one of:
Example:
bug: Login fails when the password contains special characters@owui-terminator[bot] commented on GitHub (Jul 10, 2026):
🔍 Related Issues Found
I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions:
🟣 #58 feat: reasoning content in the request
This issue asks for adding reasoning_content to requests, which is directly related to the new request about passing reasoning/thinking data back to the provider. It addresses the same class of provider-specific reasoning metadata handling.
by naqerl
🟣 #63 “'NoneType' object is not iterable” with specific model provider
Although this is a different error, it concerns provider-specific behavior with non-default OpenAI-compatible providers and shows cptr may need explicit provider support. It may help contextualize the user's complaint that provider selection and compatibility are not obvious in cptr.
by DtDEncryption
💡 If your issue is a duplicate, please close it and add any additional details to the existing issue instead.
This comment was generated automatically. React with 👍 if helpful, 👎 if not.
@kierro-max commented on GitHub (Jul 13, 2026):
I thought issue #58 addressed this but it does not seem to be the case.
@tjbck commented on GitHub (Jul 16, 2026):
@kierro-max should be addressed with the latest, testing wanted here!
@kierro-max commented on GitHub (Jul 17, 2026):
It works!
I updated to 0.9.9, went to settings, connections and changed provider type to llama.cpp, saved, and voila. Here is the conversation that proves it works:
think about a number between 10000 and 20000. I will try to guess it. do not make tool calls
The user wants me to think of a number between 10,000 and 20,000 for them to guess. They explicitly said "do not make tool calls", so I should just respond with text. I'll pick a number - let me go with 14,783. I won't reveal it; I'll just confirm that I've thought of one and invite them to start guessing.
Alright, I've thought of a number between 10,000 and 20,000. Go ahead and take your first guess! 😊
is it 15000
The user is guessing if the number is 15,000. I picked 14,783, so 15,000 is too high. I should give them a hint.
Too high! Try a lower number. 📉
is it 14000
14,000 is lower than my number (14,783), so it's too low.
Too low! Try a higher number. 📈
is it 14783
🎉 Yes! That's it! 14,783 is the number!
Nice guess - you got it in just 3 tries! Well played. 😄