AI Pro user here. Every time I boot up Antigravity, I’m getting lines like wsarecv: An established connection was aborted by the software in your host machine.
Is this a known issue? I worry this kind of behaviour’s causing issues downstream and possibly affecting quota billing.
For context, here’s the whole Output log:
————————————————————————————————-
2026-02-06 23:57:04.807 [info] 2026/02/06 23:57:04 maxprocs: Leaving GOMAXPROCS=8: CPU quota undefined
2026-02-06 23:57:04.906 [info] I0206 23:57:04.904971 11672 server.go:1192] Starting language server process with pid 11672
2026-02-06 23:57:04.906 [info] I0206 23:57:04.904971 11672 server.go:291] Setting GOMAXPROCS to 4
2026-02-06 23:57:04.911 [info] I0206 23:57:04.909300 11672 server.go:483] Language server will attempt to listen on host 127.0.0.1
2026-02-06 23:57:04.912 [info] I0206 23:57:04.911044 11672 server.go:498] Language server listening on random port at 58477 for HTTPS
2026-02-06 23:57:04.912 [info] I0206 23:57:04.911044 11672 server.go:505] Language server listening on random port at 58478 for HTTP
2026-02-06 23:57:05.255 [info] I0206 23:57:05.235285 11672 server.go:476] Created extension server client at port 58476
2026-02-06 23:57:05.396 [info] I0206 23:57:05.395757 11672 server.go:593] Using ApiServerClientV2
2026-02-06 23:57:05.441 [info] (Antigravity) 2026-02-06 23:57:05.438 [INFO]: Language server started
2026-02-06 23:57:05.499 [info] I0206 23:57:05.466748 11672 server.go:1562] initialized server successfully in 561.7767ms
2026-02-06 23:57:05.499 [info] 2026/02/06 23:57:05 http: TLS handshake error from 127.0.0.1:57408: read tcp 127.0.0.1:58477->127.0.0.1:57408: wsarecv: An established connection was aborted by the software in your host machine.
2026-02-06 23:57:05.555 [info] (Antigravity) 2026-02-06 23:57:05.553 [INFO]: LS lspClient started successfully
————————————————————————————————-
I’ve been through a whole string of attempted fixes, including:
Reset the onboarding flow
Purge local auth tokens
Connecting via mobile hotspot
Preventing restriction of local loopback connections by enabling developer mode in Windows
Checking for port conflicts
Forcing allow for AG through Windows firewall
Checking for zombie processes
Disabling IP Helper service
Checking for ghost LSP Handlers
None of these had any discernible effect.
Forcing a clean port bind by editing the shortcut with “–disable-features=NetworkServiceInProcess” seems to prevent the error at login, but it reappears as soon as a folder’s opened. I thought this was just because a new window was being opened, but disabling that behaviour in AG settings didn’t prevent it.
Is there a fix I should be looking into? I’m a bit worried about the security implications of some of these steps.