What would you like to be added?
One of the more basic things Gemini CLI could use is Antigravity's access to both Gemini models and Opus / Sonnet.
But more broadly, it should do everything Claude Code does well, and also ideally cover the other usual things that developers might need to do for which we have features in IDEs.
Side Note: I'd also suggest looking at T3 Code and Cline Kanban for reference while you're exploring the topic, since they are attempts at solving some of the problems people experience during Agentic coding.
Why is this needed?
Business reason?
Gemini CLI is close to unusable compared to Claude Code (especially on my Macbook Air) – feels like "abandoned product" levels of bad.
But more broadly, we know that Claude Code is something that users want.
Additional context
Additional Notes:
I was using Agent Manager in Antigravity recently (Antigravity is my default IDE for coding myself), but I think the Cognitive Load there ends up being higher than when you're using Codex or Claude Code... Just feels less pleasant to use... I mention that as a UX issue.
Basically this all falls under "Giving users the experiences that you know they like by seeing what's working for competitors... So that they use your products / models instead of your competitors.".
I was going to add Gemini sub-agents to my Claude Code just so I could have OpenAI, Anthropic, and Google models available in one place for ease of use where it makes sense – but I can't use Gemini CLI with my built in Google Auth to run that, because it's super slow... Even the boot-up is so slow that my mood has soured before I can try to use it for anything.
What would you like to be added?
One of the more basic things Gemini CLI could use is Antigravity's access to both Gemini models and Opus / Sonnet.
But more broadly, it should do everything Claude Code does well, and also ideally cover the other usual things that developers might need to do for which we have features in IDEs.
Side Note: I'd also suggest looking at T3 Code and Cline Kanban for reference while you're exploring the topic, since they are attempts at solving some of the problems people experience during Agentic coding.
Why is this needed?
Business reason?
Gemini CLI is close to unusable compared to Claude Code (especially on my Macbook Air) – feels like "abandoned product" levels of bad.
But more broadly, we know that Claude Code is something that users want.
Additional context
Additional Notes:
I was using Agent Manager in Antigravity recently (Antigravity is my default IDE for coding myself), but I think the Cognitive Load there ends up being higher than when you're using Codex or Claude Code... Just feels less pleasant to use... I mention that as a UX issue.
Basically this all falls under "Giving users the experiences that you know they like by seeing what's working for competitors... So that they use your products / models instead of your competitors.".
I was going to add Gemini sub-agents to my Claude Code just so I could have OpenAI, Anthropic, and Google models available in one place for ease of use where it makes sense – but I can't use Gemini CLI with my built in Google Auth to run that, because it's super slow... Even the boot-up is so slow that my mood has soured before I can try to use it for anything.