Closed (duplicate)
Project:
Drupal Canvas
Version:
1.x-dev
Component:
AI
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Issue tags:
Reporter:
Created:
10 Nov 2025 at 10:38 UTC
Updated:
31 Mar 2026 at 05:47 UTC
Jump to comment: Most recent

Comments
Comment #2
rakhimandhania commentedComment #3
akhil babuSome models, such as Claude, tend to be quite verbose during execution and generate reasoning text that reflects what they are doing. It may be beneficial to retain this behavior, as it helps users understand that the agent is actively processing, rather than simply showing a loading state. Additionally, these messages are not included in the conversation history (see the issue summary in #3555239: Canvas AI: Orchestrator missing previous conversation context).
That said, one downside is that when the AI attempts to add something to the page and fails, it may generate messages like "Let me correct the errors and try again," without clearly indicating what went wrong. In such cases, the user does not have visibility into the actual error. This issue is documented in #3568057: Chatbot responses are confusing.
Addressing this properly would likely require a broader approach, and it may be best to start with #3575780: Implement tool call visibility UI in Canvas AI.
With that in mind, IMHO, we could consider closing this issue
Comment #5
akhil babu