Summary
It's easy to forget which model is currently responding, especially when switching between Claude Code and other AI CLIs (e.g. Codex shows gpt at the bottom of its prompt). Claude Code already supports custom statusLine configs, but a sensible model-name default would be helpful out of the box.
Use case
I sometimes mix up which model is talking when I have Claude Code and Codex open at the same time. Codex's persistent "gpt" footer makes it obvious which CLI I'm in; Claude Code has no such indicator unless the user manually writes a statusLine command.
Proposed default
A built-in statusLine that displays something like:
Pulled from the same session metadata that's already available to user-defined statusLine commands (.model.display_name / .model.id). Update on /model switch (which it already does for user-defined ones).
Current workaround
For anyone who lands here, this user-side config does the job:
{
"statusLine": {
"type": "command",
"command": "python3 -c \"import json,sys; d=json.load(sys.stdin); m=d.get('model',{}); print('🤖 Claude', m.get('display_name') or m.get('id') or '?')\""
}
}(Add to ~/.claude/settings.json and restart.) Works fine, but ~99% of users won't know this exists. Hence the default suggestion.
Implementation notes (just thoughts)
- Could be the default only when no user
statusLine is configured.
- Format could match whatever convention Anthropic prefers — short ID, display name, with/without emoji.
- Cost: ~one line of code at the bottom of the prompt, identical mechanism to existing statusLine.
Summary
It's easy to forget which model is currently responding, especially when switching between Claude Code and other AI CLIs (e.g. Codex shows
gptat the bottom of its prompt). Claude Code already supports customstatusLineconfigs, but a sensible model-name default would be helpful out of the box.Use case
I sometimes mix up which model is talking when I have Claude Code and Codex open at the same time. Codex's persistent "gpt" footer makes it obvious which CLI I'm in; Claude Code has no such indicator unless the user manually writes a
statusLinecommand.Proposed default
A built-in
statusLinethat displays something like:Pulled from the same session metadata that's already available to user-defined statusLine commands (
.model.display_name/.model.id). Update on/modelswitch (which it already does for user-defined ones).Current workaround
For anyone who lands here, this user-side config does the job:
{ "statusLine": { "type": "command", "command": "python3 -c \"import json,sys; d=json.load(sys.stdin); m=d.get('model',{}); print('🤖 Claude', m.get('display_name') or m.get('id') or '?')\"" } }(Add to
~/.claude/settings.jsonand restart.) Works fine, but ~99% of users won't know this exists. Hence the default suggestion.Implementation notes (just thoughts)
statusLineis configured.