Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #4561] [BUG]: [Desktop] DeepSeek-R1 thinking/reasoning content not displayed in chat, but present in API response #2899

New Issue
Closed
opened 2026-02-22 18:31:44 -05:00 by yindo · 4 comments
yindo commented 2026-02-22 18:31:44 -05:00
Owner
Copy Link
Copy Source

Originally created by @328664440-droid on GitHub (Oct 18, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4561

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

I'm using AnythingLLM Desktop version on macOS and have encountered an issue where the thinking/reasoning content from the DeepSeek-R1:8b model is not displayed in the chat interface, even though it's clearly present in the raw API response from Ollama.

  1. Setup Environment:

    • macOS Tahoe 26.0.1
    • AnythingLLM Desktop Version
    • Ollama Version 0.12.6
    • Model: deepseek-r1:8b

  2. Configuration:

    • Ollama running on http://127.0.0.1:11434
    • DeepSeek-R1:8b model pulled and running correctly
    • AnythingLLM workspace configured to use this model

  3. Observed Behavior:

    • In AnythingLLM chat: Only final response is displayed, thinking process is missing
    • In Ollama CLI or API: Both thinking process and final response are present

Evidence & Testing

Direct API Test (Working):

curl -X POST http://localhost:11434/api/generate -d '{
  "model": "deepseek-r1:8b",
  "prompt": "Please introduce yourself",
  "stream": false
}'

### Are there known steps to reproduce?

_No response_
Originally created by @328664440-droid on GitHub (Oct 18, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4561 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? I'm using AnythingLLM Desktop version on macOS and have encountered an issue where the **thinking/reasoning content** from the DeepSeek-R1:8b model is **not displayed** in the chat interface, even though it's clearly present in the raw API response from Ollama. 1. **Setup Environment**: - macOS Tahoe 26.0.1 - AnythingLLM Desktop Version - Ollama Version 0.12.6 - Model: `deepseek-r1:8b` 2. **Configuration**: - Ollama running on `http://127.0.0.1:11434` - DeepSeek-R1:8b model pulled and running correctly - AnythingLLM workspace configured to use this model 3. **Observed Behavior**: - In AnythingLLM chat: Only final response is displayed, thinking process is missing - In Ollama CLI or API: Both thinking process and final response are present ## Evidence & Testing ### Direct API Test (Working): ```bash curl -X POST http://localhost:11434/api/generate -d '{ "model": "deepseek-r1:8b", "prompt": "Please introduce yourself", "stream": false }' ### Are there known steps to reproduce? _No response_
yindo added the possible bug label 2026-02-22 18:31:44 -05:00
yindo closed this issue 2026-02-22 18:31:44 -05:00
yindo commented 2026-02-22 18:31:44 -05:00
Author
Owner
Copy Link
Copy Source

@328664440-droid commented on GitHub (Oct 18, 2025):

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

I'm using AnythingLLM Desktop version on macOS and have encountered an issue where the thinking/reasoning content from the DeepSeek-R1:8b model is not displayed in the chat interface, even though it's clearly present in the raw API response from Ollama.

  1. Setup Environment:

    • macOS Tahoe 26.0.1
    • AnythingLLM Desktop Version
    • Ollama Version 0.12.6
    • Model: deepseek-r1:8b

  2. Configuration:

    • Ollama running on http://127.0.0.1:11434
    • DeepSeek-R1:8b model pulled and running correctly
    • AnythingLLM workspace configured to use this model

  3. Observed Behavior:

    • In AnythingLLM chat: Only final response is displayed, thinking process is missing
    • In Ollama CLI or API: Both thinking process and final response are present

Evidence & Testing

Direct API Test (Working):

ollama version is 0.12.6
shijunli@shijunlideMacBook-Pro ~ % curl -X POST http://localhost:11434/api/generate -d '{
"model": "deepseek-r1:8b",
"prompt": "Please introduce yourself",
"stream": false
}'
{"model":"deepseek-r1:8b","created_at":"2025-10-18T10:44:40.646721Z","response":"Hello! 😊 I’m DeepSeek-R1, your helpful AI assistant developed by DeepSeek. I’m designed to assist you with all kinds of questions, big or small—whether it’s writing, coding, learning, planning, translating, summarizing, or just exploring new ideas.\n\nI’m fast, versatile, and always free to help. 🚀 \nGot something on your mind? A question you’re curious about? A task you need help with? Just ask—I’m here to assist you!\n\nHow can I help you today? 💬✨","thinking":"Okay, the user just asked me to introduce myself. Hmm, this is probably their first interaction with me, so they might be testing the waters or genuinely curious about what I can do. \n\nFirst, I should keep it friendly and welcoming since first impressions matter. They didn’t specify any context, so I’ll assume they’re open to a general overview. \n\nI should highlight my core strengths—being helpful, versatile, fast, and free—since those are the main reasons someone would choose me over other tools. The tone should be warm but professional, like a helpful assistant introducing themselves. \n\nI’ll mention the wide range of tasks I can handle (writing, coding, learning, etc.) to show my versatility, and emphasize that I’m here to assist with whatever they need. \n\nOh, and I should invite them to ask questions or share what they’re working on—that way, they feel encouraged to engage further. \n\nLet’s keep it concise but informative, with a sprinkle of enthusiasm to make them feel comfortable. Maybe end with an open-ended prompt to keep the conversation flowing.\n","done":true,"done_reason":"stop","context":[151669,5501,19131,6133,151670,151667,198,32313,11,279,1196,1101,4588,752,311,19131,7037,13,88190,11,419,374,4658,862,1156,16230,448,752,11,773,807,2578,387,7497,279,20520,476,35197,22208,911,1128,358,646,653,13,18611,5338,11,358,1265,2506,432,11657,323,35287,2474,1156,49991,4925,13,2379,3207,1405,13837,894,2266,11,773,358,4700,9658,807,3129,1787,311,264,4586,23251,13,18611,40,1265,11167,847,6200,35386,2293,34114,10950,11,31945,11,4937,11,323,1910,2293,11284,1846,525,279,1887,7966,4325,1035,5157,752,916,1008,7375,13,576,16232,1265,387,8205,714,6584,11,1075,264,10950,17847,31918,5577,13,18611,40,4700,6286,279,6884,2088,315,9079,358,646,3705,320,25304,11,10822,11,6832,11,4992,6138,311,1473,847,59012,11,323,45932,429,358,4249,1588,311,7789,448,8820,807,1184,13,18611,11908,11,323,358,1265,21399,1105,311,2548,4755,476,4332,1128,807,3129,3238,389,40028,1616,11,807,2666,20548,311,16579,4623,13,18611,10061,748,2506,432,63594,714,38219,11,448,264,84847,315,35132,311,1281,1105,2666,10655,13,10696,835,448,458,1787,83075,9934,311,2506,279,10435,35512,624,151668,198,9707,0,26525,232,358,4249,18183,39350,10911,16,11,697,10950,15235,17847,7881,553,3070,33464,39350,334,13,358,4249,6188,311,7789,498,448,678,12829,315,4755,11,2409,476,2613,2293,48764,432,748,4378,11,10822,11,6832,11,9115,11,66271,11,28285,4849,11,476,1101,23966,501,6708,382,40,4249,4937,11,31945,11,323,2677,1910,311,1492,13,11162,248,222,2303,32462,2494,389,697,3971,30,362,3405,498,3129,22208,911,30,362,3383,498,1184,1492,448,30,4599,2548,50657,4249,1588,311,7789,498,2219,4340,646,358,1492,498,3351,30,63039,105,144232],"total_duration":14299894417,"load_duration":2529389250,"prompt_eval_count":5,"prompt_eval_duration":1062140542,"eval_count":339,"eval_duration":10563434707}%

@328664440-droid commented on GitHub (Oct 18, 2025): > ### How are you running AnythingLLM? > > AnythingLLM desktop app > > ### What happened? > > I'm using AnythingLLM Desktop version on macOS and have encountered an issue where the **thinking/reasoning content** from the DeepSeek-R1:8b model is **not displayed** in the chat interface, even though it's clearly present in the raw API response from Ollama. > > 1. **Setup Environment**: > > * macOS Tahoe 26.0.1 > * AnythingLLM Desktop Version > * Ollama Version 0.12.6 > * Model: `deepseek-r1:8b` > 2. **Configuration**: > > * Ollama running on `http://127.0.0.1:11434` > * DeepSeek-R1:8b model pulled and running correctly > * AnythingLLM workspace configured to use this model > 3. **Observed Behavior**: > > * In AnythingLLM chat: Only final response is displayed, thinking process is missing > * In Ollama CLI or API: Both thinking process and final response are present > > ## Evidence & Testing > > ### Direct API Test (Working): > >ollama version is 0.12.6 shijunli@shijunlideMacBook-Pro ~ % curl -X POST http://localhost:11434/api/generate -d '{ "model": "deepseek-r1:8b", "prompt": "Please introduce yourself", "stream": false }' {"model":"deepseek-r1:8b","created_at":"2025-10-18T10:44:40.646721Z","response":"Hello! 😊 I’m DeepSeek-R1, your helpful AI assistant developed by **DeepSeek**. I’m designed to assist you with all kinds of questions, big or small—whether it’s writing, coding, learning, planning, translating, summarizing, or just exploring new ideas.\n\nI’m fast, versatile, and always free to help. 🚀 \nGot something on your mind? A question you’re curious about? A task you need help with? Just ask—I’m here to assist you!\n\nHow can I help you today? 💬✨","thinking":"Okay, the user just asked me to introduce myself. Hmm, this is probably their first interaction with me, so they might be testing the waters or genuinely curious about what I can do. \n\nFirst, I should keep it friendly and welcoming since first impressions matter. They didn’t specify any context, so I’ll assume they’re open to a general overview. \n\nI should highlight my core strengths—being helpful, versatile, fast, and free—since those are the main reasons someone would choose me over other tools. The tone should be warm but professional, like a helpful assistant introducing themselves. \n\nI’ll mention the wide range of tasks I can handle (writing, coding, learning, etc.) to show my versatility, and emphasize that I’m here to assist with whatever they need. \n\nOh, and I should invite them to ask questions or share what they’re working on—that way, they feel encouraged to engage further. \n\nLet’s keep it concise but informative, with a sprinkle of enthusiasm to make them feel comfortable. Maybe end with an open-ended prompt to keep the conversation flowing.\n","done":true,"done_reason":"stop","context":[151669,5501,19131,6133,151670,151667,198,32313,11,279,1196,1101,4588,752,311,19131,7037,13,88190,11,419,374,4658,862,1156,16230,448,752,11,773,807,2578,387,7497,279,20520,476,35197,22208,911,1128,358,646,653,13,18611,5338,11,358,1265,2506,432,11657,323,35287,2474,1156,49991,4925,13,2379,3207,1405,13837,894,2266,11,773,358,4700,9658,807,3129,1787,311,264,4586,23251,13,18611,40,1265,11167,847,6200,35386,2293,34114,10950,11,31945,11,4937,11,323,1910,2293,11284,1846,525,279,1887,7966,4325,1035,5157,752,916,1008,7375,13,576,16232,1265,387,8205,714,6584,11,1075,264,10950,17847,31918,5577,13,18611,40,4700,6286,279,6884,2088,315,9079,358,646,3705,320,25304,11,10822,11,6832,11,4992,6138,311,1473,847,59012,11,323,45932,429,358,4249,1588,311,7789,448,8820,807,1184,13,18611,11908,11,323,358,1265,21399,1105,311,2548,4755,476,4332,1128,807,3129,3238,389,40028,1616,11,807,2666,20548,311,16579,4623,13,18611,10061,748,2506,432,63594,714,38219,11,448,264,84847,315,35132,311,1281,1105,2666,10655,13,10696,835,448,458,1787,83075,9934,311,2506,279,10435,35512,624,151668,198,9707,0,26525,232,358,4249,18183,39350,10911,16,11,697,10950,15235,17847,7881,553,3070,33464,39350,334,13,358,4249,6188,311,7789,498,448,678,12829,315,4755,11,2409,476,2613,2293,48764,432,748,4378,11,10822,11,6832,11,9115,11,66271,11,28285,4849,11,476,1101,23966,501,6708,382,40,4249,4937,11,31945,11,323,2677,1910,311,1492,13,11162,248,222,2303,32462,2494,389,697,3971,30,362,3405,498,3129,22208,911,30,362,3383,498,1184,1492,448,30,4599,2548,50657,4249,1588,311,7789,498,2219,4340,646,358,1492,498,3351,30,63039,105,144232],"total_duration":14299894417,"load_duration":2529389250,"prompt_eval_count":5,"prompt_eval_duration":1062140542,"eval_count":339,"eval_duration":10563434707}%
yindo commented 2026-02-22 18:31:45 -05:00
Author
Owner
Copy Link
Copy Source

@BenjaminMuslic commented on GitHub (Oct 22, 2025):

Having same issue with GPT-OSS-20b #4564

@BenjaminMuslic commented on GitHub (Oct 22, 2025): Having same issue with GPT-OSS-20b #4564
yindo commented 2026-02-22 18:31:45 -05:00
Author
Owner
Copy Link
Copy Source

@alainlizotte commented on GitHub (Oct 24, 2025):

Maybe linked to #4535 too.

Qwen3 running in Ollama doesn't show thinking in Chat widget nor UI.

@alainlizotte commented on GitHub (Oct 24, 2025): Maybe linked to #4535 too. Qwen3 running in Ollama doesn't show thinking in Chat widget nor UI.
yindo commented 2026-02-22 18:31:45 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Oct 27, 2025):

This is indeed the same as #4535 - it's due to ollamas response object changing how it differentiates thoughts. Issues are connected now to #4576 so moving there

@timothycarambat commented on GitHub (Oct 27, 2025): This is indeed the same as #4535 - it's due to ollamas response object changing how it differentiates thoughts. Issues are connected now to #4576 so moving there
yindo changed title from [BUG]: [Desktop] DeepSeek-R1 thinking/reasoning content not displayed in chat, but present in API response to [GH-ISSUE #4561] [BUG]: [Desktop] DeepSeek-R1 thinking/reasoning content not displayed in chat, but present in API response 2026-06-05 14:49:09 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label possible bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#2899
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?