← Back to AI Insights
Gemini Executive Synthesis

Control and display of AI model reasoning/thinking process output.

Technical Positioning
User experience, output filtering, configuration management, AI transparency.
SaaS Insight & Market Implications
The persistent display of raw `` tags despite multiple attempts to disable reasoning output indicates a critical failure in the UI's content filtering and configuration management. This directly impacts user experience by cluttering responses with internal model processing, undermining the intended clarity and professionalism of the AI agent's output. For a B2B SaaS product, such a bug suggests a lack of robust control over AI model verbosity, which is essential for tailoring agent interactions to specific business contexts and user preferences. This issue requires immediate resolution to ensure a polished, configurable user interface.
Proprietary Technical Taxonomy
raw thinking tags reasoning hide display.show_reasoning: false verbose off Show Reasoning (Show model thinking process) config.yaml Hermes Agent version Model used: Minimax M2.7

Raw Developer Origin & Technical Request

Source Icon GitHub Issue Apr 17, 2026
Repo: EKKOLearnAI/hermes-web-ui
关闭不了

### Description

I'm still seeing **raw thinking tags** (e.g. ``, `` in the response.

Even after using:

- `/reasoning hide`
- Setting `display.show_reasoning: false` in `~/.hermes/config.yaml`
- `/verbose off`
- toggled OFF : `Show Reasoning (Show model thinking process)`

the tags still appear.

### Steps to Reproduce
1. Run `hermes`
2. Send any message that triggers reasoning
3. Observe raw `...` (or similar) tags in the output instead of being stripped or hidden.

### Expected Behavior
Thinking/reasoning content should be completely hidden (when `show_reasoning: false` or `/reasoning hide` is used) or properly stripped so no raw tags appear in the final message.

### Actual Behavior
Raw tags are printed visibly in the UI.

### Environment
- Hermes Agent version: `Hermes Agent v0.10.0 (2026.4.16)`
- Model used: Minimax M2.7
- Installation method: bash

Developer Debate & Comments

No active discussions extracted for this entry yet.

Adjacent Repository Pain Points

Other highly discussed features and pain points extracted from EKKOLearnAI/hermes-web-ui.

Extracted Positioning
Visibility and accessibility of the login token (authentication token) for the Hermes web UI.
User authentication, ease of access, security management, operational transparency.
Top Replies
anyu5733 • Apr 18, 2026
遇到同样错误 docker-compose logs -f hermes-webui hermes-webui | ⚠ Gateway not running, starting... hermes-webui | ✓ Gateway started (PID: 20) hermes-webui | [Terminal] WebSocket ready at /terminal ...
windfall4us • Apr 18, 2026
建议这个登陆令牌。能不能显示在web端。换个电脑就很难找到。
kaka77 • Apr 18, 2026
+1 ，日志中没看到令牌。
Extracted Positioning
Bug in displaying multiple user profiles; only 'default' is shown.
Multi-profile support, user interface accuracy, configuration management.
Top Replies
ZiDuNet • Apr 18, 2026
排查你的日志吧
yinoliver-MI • Apr 19, 2026
我问Hermes（Minimax 2.7 cn模型），它告诉我只支持 单用户，多Provider，多模型模式。 原来是可以增加 “用户”的。
yinoliver-MI • Apr 19, 2026
创建失败！
Extracted Positioning
Core AI chat functionality (message sending/receiving, model listing, provider integration).
Functional reliability, AI model integration, logging and debugging.
Extracted Positioning
UI accessibility features (font size adjustment, dark mode contrast).
User experience, accessibility, developer-friendly design, long-term usability.
Extracted Positioning
Feature integration from other Hermes UI projects (workspace, login authentication key management, configuration file management, key management).
Feature completeness, user configurability, security management, competitive differentiation.

Frequently Asked Questions

Market intelligence mapped to Control and display of AI model reasoning/thinking process output..

What is the technical positioning of Control and display of AI model reasoning/thinking process output.?
Based on our AI analysis of the original developer request, its primary technical positioning is: User experience, output filtering, configuration management, AI transparency.
Are engineers actively discussing Control and display of AI model reasoning/thinking process output.?
Yes, we have tracked 2 direct responses and active debates regarding this specific topic originating from GitHub Issue.
What architecture is tied to Control and display of AI model reasoning/thinking process output.?
Our proprietary extraction maps Control and display of AI model reasoning/thinking process output. to adjacent architectural concepts including raw thinking tags, reasoning hide, display.show_reasoning: false, verbose off.

Engagement Signals

2
Replies
open
Issue Status

Cross-Market Term Frequency

Quantifies the cross-market adoption of foundational terms like raw thinking tags and reasoning hide by tracking occurrence frequency across active SaaS architectures and enterprise developer debates.