Describe the feature or problem you'd like to solve
The CLI adds an artificial delay when rendering streamed output, producing a typewriter-like effect. For users who read completed blocks rather than watching tokens appear, this slows down the perceived response time without adding value. There is no setting to disable or reduce this behaviour.
Proposed solution
Add a setting (e.g. in settings.json or via a --no-animate flag) to render streamed output as fast as it arrives from the API, without an artificial per-token or per-chunk delay. This would benefit users on fast connections who prefer to read complete output, and would improve the experience when using screen readers or piping output to other tools.
Example prompts or workflows
- settings.json:
{ "streamingDelay": 0 } or { "instantOutput": true }
- CLI flag:
copilot --no-animate
- A
/speed or /animation slash command to toggle during a session
Additional context
Running on Linux (Debian via WSL2) with a wired connection. The delay is most noticeable on longer responses where waiting for the full render adds significant wall-clock time compared to the actual API response time.
Describe the feature or problem you'd like to solve
The CLI adds an artificial delay when rendering streamed output, producing a typewriter-like effect. For users who read completed blocks rather than watching tokens appear, this slows down the perceived response time without adding value. There is no setting to disable or reduce this behaviour.
Proposed solution
Add a setting (e.g. in settings.json or via a
--no-animateflag) to render streamed output as fast as it arrives from the API, without an artificial per-token or per-chunk delay. This would benefit users on fast connections who prefer to read complete output, and would improve the experience when using screen readers or piping output to other tools.Example prompts or workflows
{ "streamingDelay": 0 }or{ "instantOutput": true }copilot --no-animate/speedor/animationslash command to toggle during a sessionAdditional context
Running on Linux (Debian via WSL2) with a wired connection. The delay is most noticeable on longer responses where waiting for the full render adds significant wall-clock time compared to the actual API response time.