Repository navigation
[Bug]: Extreme energy consumption, 12hr power 136x higher than comparable Electron apps #3143
Description
Activity
that number includes all subprocesses spawned by agents as well, if your agents run heavy
tsgoorvitestprocesses, those all add up to that number. so you can't just say "it's heavy compared to other electron apps"Thanks for the context, Julius. In this case though, I didn't have any processes running and was doing my work in the Claude Code app, not through T3 Code. The app was essentially just open and idle. So there were no agent-spawned subprocesses.
OK. Thanks for confirming. I do have work in progress to reduce background activity whilst idle (#2679). Will hopefully land that within the next few days
Reacted by ILIA and gdfelttiHere is another example to reference on Windows, of very high power usage, (all sessions are completely idle).
They also don't seem to be grouped under the T3 Code process, so that's interesting to me.
If they are detached those are often spawned by Codex itself. They have a bug where the agent spawns subprocesses detached. I've raised this to their team and they are looking into it
Reacted by gdfeltticool, in this case I was using Claude, would that be a similar issue?
I don't know. What command was it running? Can you get some more info other than it's a generic node process?
The only commands that I saw where some git analytics for the latest commit, and then it ran TSC presumably, and then it did some reads, but at least one node process seemed to appear almost immediately after I kicked off the session. I couldn't find any node related commands specifically. which was what lead me to believe it was something internal to T3.
(The processes seemed to increase with each interaction I had in the session).
I don't know. What command was it running? Can you get some more info other than it's a generic node process?
I was trying hard to determine where the node processes came from, I think I'll have to find time to more properly debug where processes are coming from, task-manager, or resource monitor don't have any kind of "parent" tracking as to where the process came from.
If you don't have time to recreate the bug, I'll find time in the next few days to try and find the actual source creating the processes.
Okay I think I figured it out, (at least for my case). I had the 'Official' Chrome Devtools mcp (Plugin) enabled in Claude Code, but when I disabled it my processes went away while using T3 (even idle), and when I enabled it again, they re-appeared when starting T3.
Yup this is a huge issue. Screenshot for reference:
Note - all of Warp, Zed and T3Code had a similar number of threads going (<4 each) - I was testing the workflows on all of them.
While I like the T3 flow the best, the energy consumption is a deal breaker. I didn't take a screenshot of the thermals and CPU at the time but both were over 100 (both around 20 without it).
Before submitting
Area
apps/desktop
Steps to reproduce
Expected behavior
T3 Code’s 12-hour cumulative power usage should be in the same ballpark as other Electron/Chromium-based apps (roughly 50-300 units).
Actual behavior
T3 Code reports a 12hr Power value of 12,295.04, higher than any comparable apps running at the same time:
That’s ~136x Claude and ~354x Dia. The current Energy Impact (11.6) is modest, which suggests the drain may be intermittent through a possible background process spiking periodically rather than sustained high usage.
Impact
Major degradation or frequent failure
Version or commit
No response
Environment
No response
Logs or stack traces
Screenshots, recordings, or supporting files
Screenshot 2026-06-18 at 14.23.51.png
Workaround
No response