Kilo DevRel here. We built the Kilo CLI because other CLI tools either lock you into one AI coding model, or make you jump through hoops to switch. We wanted agentic coding in the terminal with the same flexibility you'd get from the Kilo IDE setup.
What makes it different:
- 500+ models (including everyone’s favorites from Claude, Gemini, OpenAI). Switch anytime
- 5 specialized modes: Code, Ask, Debug, Architect, and Orchestrator (breaks down tasks and delegates to the others)
- Parallel agents: spin up multiple instances on the same repo, each on its own branch
- Session continuity: start in CLI, pick up in Kilo VS Code or cloud. Context travels with you
--parallel --auto for running multiple autonomous agents simultaneously is a favorite. It’s the same engine as our IDE extensions, not a stripped-down version.
Install: npm install -g @kilocode/cli
Also: open source, and contributions are welcome. Would love feedback on what's missing from your terminal workflow.
1000% this. Today LLMs are like enthusiastic, energetic, over-confident, well-read junior engineers.
Does it take effort to work with them and get them to be effective in your code base? Yes. But is there a way to lead them in such a way that your "team" (you in this case) gets more done? Yes.
But it does take effort. That's why I love "vibe engineering" as a term because the engineering (or "senior" or "lead" engineering) is STILL what we are doing.
I would love to know if and how you disagree. I am (I thought obviously) making a jump here from her desire to make FLOW-MATIC human-readable. Also, I don't believe that AI will take developers's jobs, but I do believe a developer using AI will replace ones not using it...so I'm intentionally provocative there, yes.
It's probably true in our industry too - I'm sure that when people started using programming languages instead of writing machine code people looked down on them too.
This will be another abstraction layer that MANY people will use and be able to accomplish things that would have been impossible to do in a reasonable amount of time in machine code.
Author and Kilo Code team member here - this is a much better explanation of what I mean. And honestly, it's a quick phrase I've been using that is shorthand really for THIS much better-written take.
Yes - with our built-in provider, we provide all the models that OpenRouter provides but without OpenRouter's 5% markup. We provide them at cost (the AI provider cost)
Our plan is to be a superset of Cline and Roo's features (we already have all the major features from both) [0]
We also have our own provider, which means no need to bring your own API keys (you can if you like, but it is batteries included by default) and we're not charging anything on top of the API pricing. Instead of monetizing on individual developers, we want it to be free for them and make money eventually off enterprise contracts [1]
Marathon Crater isn’t one of the long‐known,
named lunar features from 19th-century telescopic maps –
it was first singled out and informally “discovered”
by NASA in the mid-1960s when Lunar Orbiter imagery
and, more definitively, the Surveyor 5 landing
(September 1967) identified it. So, very roughly,
think “around 1966–67.”
You’re not wrong but also this does raise a central question that I think is super un-considered in this whole MCP thing: how are we handling identity in those contexts.
If anything we should be more concerned so it that because of the power that it can hand over to agents.