You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Do we have an issue already for Kitty image performance issues?
Not sure how much this is relevant, and whether it is good/bad to report this here, but I've recently made a comparison between Kitty vs Ghostty by running it side-by-side (same flags / env), and the difference is clearly visible (doesn't really require nerd stats to compare against): https://www.youtube.com/watch?v=VYRJ4h1JnLU
This is on NVIDIA proprietary userspace and open-source kernel drivers. Maybe AMD / Nouveau behaves differently with Ghostty, having some data on that for at least informative purposes would be great.
I'm going to close this meta issue. We now faithfully implement almost all of the Kitty images spec with no known issues in what we support. We specifically do NOT implement: animation and parent-relative placements. But those can remain their own issues.
I've personally re-read the full spec plus Kitty's graphics.c to compare. I've also spent two full Fable plus GPT-5.6 subs over the past 3 weeks analyzing and reanalyzing the spec + reference implementation versus us to find mismatches (including those not in spec but upstream in Kitty). At this point they've come back clean multiple times with different wording after fixing a couple dozen issues: https://github.com/ghostty-org/ghostty/pulls?q=is%3Apr+kitty+is%3Aclosed+author%3Amitchellh+updated%3A%3E2026-08-10
Activity
jcollie commentedon Aug 18, 2025
Do we have an issue already for Kitty image performance issues?
pluiedev commentedon Aug 18, 2025
I don't think so... we need to document it anyways (incl. mpv flags and so on)
SpacingBat3 commentedon Jun 3, 2026
Not sure how much this is relevant, and whether it is good/bad to report this here, but I've recently made a comparison between Kitty vs Ghostty by running it side-by-side (same flags / env), and the difference is clearly visible (doesn't really require nerd stats to compare against):
https://www.youtube.com/watch?v=VYRJ4h1JnLU
This is on NVIDIA proprietary userspace and open-source kernel drivers. Maybe AMD / Nouveau behaves differently with Ghostty, having some data on that for at least informative purposes would be great.
mitchellh commentedon Aug 20, 2026
I'm going to close this meta issue. We now faithfully implement almost all of the Kitty images spec with no known issues in what we support. We specifically do NOT implement: animation and parent-relative placements. But those can remain their own issues.
I've personally re-read the full spec plus Kitty's graphics.c to compare. I've also spent two full Fable plus GPT-5.6 subs over the past 3 weeks analyzing and reanalyzing the spec + reference implementation versus us to find mismatches (including those not in spec but upstream in Kitty). At this point they've come back clean multiple times with different wording after fixing a couple dozen issues: https://github.com/ghostty-org/ghostty/pulls?q=is%3Apr+kitty+is%3Aclosed+author%3Amitchellh+updated%3A%3E2026-08-10