to interact with your coding agents is very important. There is a large variety of options out there that you can try, and they all impact how effectively you’re able to interact with your agents.
I’ve spent a lot of time testing out different platforms to orchestrate coding agents, and in this article, I’ll give my opinion on some different tools and how you can find the tool that works best for you.
I think it’s worth noting that the optimal tool will vary from person to person. A tool that I like might be a tool that you dislike, and vice versa.
Thus, I’ll also cover how to find out for yourself what works for you in this article
I’m not sponsored by any of the tools I cover in this article. These are simply tools that I personally find work well.
First of all, I always like to cover the why in each article that I write. Here I wanted to highlight why you should spend time at all testing out different interfaces to interact with your coding agents.
The main reason you should spend time doing this is that it can be more productive in the long run. If you can find an interface that works super well for you, it will make it easier for you to interact with your coding agents, and you’ll become more productive.
I strongly believe there is a pretty stark difference between the tool I like the least and the tool that I like the most, and it strongly impacts how much I’m able to do.
One of the main issues I find with some tools is that they are not easy to have a full overview of all the agents that I’m running at the same time. In this case, I have to spend time manually catching up or finding relevant sessions that I want to work on, which is, of course, wasted time that could be saved using a better tool.
In this section, I’ll go through some different interfaces for coding agents, covering some different applications.
The applications that I’ve tried for coding agents are:
Now I’ll cover each of them simply with some pros and cons.
This was my go-to terminal for quite a while, and I think Warp is overall pretty good. It’s a simple terminal with some AI features, such as autocomplete when you write commands, and you can also easily run AI models inside Warp. They also have automatic naming of different sessions and allow you to split tabs. One downside I noticed with Warp is that it is a very plain terminal and doesn’t really have any added features, and I did experience it lagging a bit even though my computer was not lagging; I have not experienced issues like this with other applications.
Conductor is a very nice beginner-friendly terminal. It’s a very clean setup. I think the way they organize the different agents that you’re running is very good. They organize it into backlog, in progress, in review, and done, and then allow you to archive tabs and bring them back if you want to. The only downside I noticed in Conductor is actually two things. It doesn’t allow you to split tabs, which is very bad, of course. And secondly, it doesn’t have complete feature parity with the coding agents that I prefer. One way I noticed this is that it doesn’t have the slash goal command available for Claude Code sessions. It only has it for Codex sessions.
This was a deal-breaker for me, so I had to stop using Conductor just because of that.
Emdash is the current application that I’m using, and it’s my favorite at the moment. Emdash basically has everything Conductor has, but it also has feature parity with all CLI coding agents because it literally runs the terminal inside the application, and it allows you to split panes.
The only downside, the way I see it, is that Emdash doesn’t have the equally good organization of tabs on the left hand side where you see all of your coding agents, i.e., it doesn’t have the backlog/ in progress / in review / done organization that you get in Conductor, but it still has a relatively good way of organizing the different agent tabs.
iTerm2 is also an okay option; it’s just a basic terminal where you can run your coding agents. However, I see no reason to; I stopped using a plain terminal for a reason: it’s just that while they work fine because coding agents are accessible through a CLI, there are no real features, and it doesn’t really organize the tabs that nicely for you. So I stopped using iTerm2.
The Claude Code application is quite good and very beginner-friendly. However, again, I don’t like the way they organize the agent tabs on the left-hand side and the sessions that are running. I just believe that there are applications that organize the different tabs way better, such as Emdash and Conductor.
It’s worth noting that I do use the Claude Code application through my phone because I have remote control on. And this works very well, and I really like the Claude Code application in this way.
Codex application is basically the same as the Claude Code application, and I see no real differences in either the pros or the cons.
Omnara was marketed as kind of an application that syncs really well between your phone and your computer. While it did sync the phone and PC well, I found the overall app a bit clunky to use, and I didn’t find it that intuitive. Seems to me that the design wasn’t fully up to date, and that both Conductor and Emdash were just way simpler, easier, and more intuitive to use.
Cursor was the way I started with Agentic coding and tab completion in VS Code a bit after the release of ChatGPT. Cursor is really good. However, if you only use Cursor, it gets quite expensive because a lot of the cost is usage-based and not covered by the subscription, like in Claude Code or Codex. And also, I don’t believe I need to look at any code anymore, so I don’t really see a reason to use Cursor. You can of course use pure agent view in Cursor, but at that point it’s just way pricier than alternatives such as Claude Code or Codex.
Note that there are a lot of other interfaces out there, and I haven’t tested everything. It seems like there are new agent interfaces popping up basically every day. And it can, of course, be too much to test out yourself. What I urge you to do is pay some attention to the market of agent interfaces. If you see something interesting, quickly review their website, see if it fits your needs, and if it does, then you should try it out for a bit. I believe you’ll notice pretty quickly if a coding agent interface is worth using or not, say within twenty minutes.
Now we need to move into how you can find your optimal coding interface. I’ve covered a few things that you should take into account and think about when choosing an agent interface:
And that’s it. Overall, I think you should, when testing out coding agent interfaces, think about what matters to you and what’s important to you. For some people, the most important thing is that it’s super simple to use and intuitive, while feature parity doesn’t matter that much. If so, I would say Conductor is a great choice. Others prefer a lightweight terminal above all else. If so, Warp or iTerm2.
In the end, you yourself have the best taste for what you prefer, and you should simply try out the things you find interesting and see if it works for you better than the current interface that you’re using.
In this article, I covered some different coding agent interfaces. I started off talking about why you should care about it and how much it can impact how productive you are as a programmer. I’ve covered a lot of different coding agent interfaces that I’ve tried myself, covering the pros and the cons. In the end, the optimal coding agent interface for you is whatever you like using the most. Different people have different preferences on what’s important when interacting with coding agents, and you should try out different options for yourself. I covered and mentioned quite a few examples that you can look at if you find them interesting. And I urge you to try out a few different tools and spend some time on this because of how big an impact it can have on how productive you are when working with your coding agents. It’s worth investing some time to figure out the optimal coding setup for you.
👉 My free eBook and Webinar:
🚀 10x Your Engineering with LLMs (Free 3-Day Email Course)
📚 Get my free Vision Language Models ebook
💻 My webinar on Vision Language Models
👉 Find me on socials:
💌 Substack
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。