[FEATURE] Stream thoughts word-by-word during task execution instead of displaying all at completion
Preflight Checklist
- [x] I have searched existing requests and this feature hasn't been requested yet
- [x] This is a single feature request (not multiple features)
Problem Statement
I deeply miss the old word-by-word stream of thoughts.
With the new verbose setting, all of Claude's thoughts can be made visible, but they only appear once Claude moves on to a next step, i.e. calling a tool. In my workflow, this mostly defeats the purpose of seeing thoughts.
In previous versions of Claude Code, I monitored its stream of thoughts during the task to assess whether I like the approach it's taking, or intervene if it's confused, or making assumptions about intent which I didn't want it to make.
In the current version, I can only sit and wait for minutes and hope it's doing the right thing.
If nothing else, watching it think was much more entertaining than just staring at a "thinking..." timer.
Proposed Solution
Rather than high verbosity simply controlling whether you get a big block of thoughts at the end, thoughts should be shown word-by-word as Claude works.
Alternative Solutions
_No response_
Priority
High - Significant impact on productivity
Feature Category
Configuration and settings
Use Case Example
_No response_
Additional Context
_No response_
This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗