[BUG/FEATURE]: Desktop GUI - Thinking process disappears and cannot be viewed in detail #6405

Open
opened 2026-02-16 18:04:04 -05:00 by yindo · 0 comments
Owner

Originally created by @openAgi2 on GitHub (Jan 15, 2026).

Originally assigned to: @adamdotdevin on GitHub.

Problem Description

In the Desktop GUI (not TUI), there are two issues with the thinking/reasoning display:

Issue 1: Thinking content disappears after completion

  • When using a thinking model (e.g., Claude with extended thinking), the thinking process is visible during streaming
  • Once thinking completes, the content disappears from the GUI entirely
  • Users cannot review the model's reasoning process after the fact

Issue 2: No detailed thinking view

  • Even when manually expanding the thinking block, it does not show the full/detailed thinking content
  • The detailed thought chain that was visible during streaming is lost

Expected Behavior

  1. Thinking blocks should persist after completion (like in TUI with /thinking toggle)
  2. Users should be able to expand thinking blocks to see the full detailed reasoning
  3. This setting should be persistable (remember user preference across sessions)

Current Behavior

  • Thinking content vanishes after the model finishes reasoning
  • No way to review the reasoning process
  • No persistent setting to control thinking visibility in Desktop GUI

Environment

  • OpenCode Desktop App (latest)
  • Affects all thinking-enabled models

Reference

  • TUI has /thinking command that works correctly with kv.signal("thinking_visibility", true) persistence
  • Desktop GUI lacks equivalent functionality

Suggested Implementation

  • Add a toggle in Desktop GUI settings/toolbar to control thinking visibility
  • Persist the thinking content in the message history
  • Store user preference for default expanded/collapsed state
Originally created by @openAgi2 on GitHub (Jan 15, 2026). Originally assigned to: @adamdotdevin on GitHub. ## Problem Description In the Desktop GUI (not TUI), there are two issues with the thinking/reasoning display: ### Issue 1: Thinking content disappears after completion - When using a thinking model (e.g., Claude with extended thinking), the thinking process is visible during streaming - **Once thinking completes, the content disappears from the GUI entirely** - Users cannot review the model's reasoning process after the fact ### Issue 2: No detailed thinking view - Even when manually expanding the thinking block, it does not show the full/detailed thinking content - The detailed thought chain that was visible during streaming is lost ### Expected Behavior 1. Thinking blocks should **persist** after completion (like in TUI with `/thinking` toggle) 2. Users should be able to **expand** thinking blocks to see the **full detailed reasoning** 3. This setting should be **persistable** (remember user preference across sessions) ### Current Behavior - Thinking content vanishes after the model finishes reasoning - No way to review the reasoning process - No persistent setting to control thinking visibility in Desktop GUI ### Environment - OpenCode Desktop App (latest) - Affects all thinking-enabled models ### Reference - TUI has `/thinking` command that works correctly with `kv.signal("thinking_visibility", true)` persistence - Desktop GUI lacks equivalent functionality ### Suggested Implementation - Add a toggle in Desktop GUI settings/toolbar to control thinking visibility - Persist the thinking content in the message history - Store user preference for default expanded/collapsed state
yindo added the web label 2026-02-16 18:04:04 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#6405