working with your coding agents is very important. There are many different options to try, and they all have an impact on how effectively you can work with your agents.
I've spent a lot of time exploring different platforms to integrate coding agents, and in this article, I'll give my opinion on some of the different tools and how to find the tool that works best for you.
I think it's important to note that the right tool will vary from person to person. A tool I like may be a tool you don't like, and vice versa.
Therefore, I will also cover how you can find out for yourself what works for you in this article
I am not sponsored by any of the tools I include in this article. These are just tools that I find work well.
Why waste time finding the right agent interface?
First, I always like to cover the why in each article I write. Here I wanted to highlight why you should always spend time exploring interfaces to communicate 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 best for you, it will make it easier for you to communicate with your coding agents, and you will be more productive.
I'm a firm believer that there's a huge difference between a tool I like a little and a tool I like a lot, and it really affects how much I'm able to do.
One of the main problems I find with other tools is that it is not easy to have a complete view of all the agents I use at the same time. In this case, I have to spend time manually catching or finding the right moments that I want to work on, which is, of course, wasted time that could be saved using a better tool.
Encoding agent interfaces
In this section, I will go through some different areas of coding agents, covering different applications.
The apps I tried for coding agents are:
- Wap
- The operator
- Emdash
- Term 2
- Claude code request
- Codex Application
- Omnara
- Cursor
Now I will cover each of them with pros and cons.
Wap
This has been my go-to terminal for a long time, and I think Warp is great. It's a simple terminal with AI features, such as auto-completion when typing commands, and you can easily use AI models within Warp. They also have automatic naming of separate sessions and allow you to split tabs. One negative thing I noticed about Warp is that it's a plain terminal and doesn't have any extra features, and I've noticed that it lags a bit even though my computer isn't lagging; I haven't encountered problems like this in other applications.
The operator
Conductor is an excellent terminal for beginners. Very clean setup. I think the way they organize the different agents that they manage is very good. They organize it into backlog, in progress, review, and done, and then allow you to archive tabs and restore them if you want. The only fault I've noticed with Conductor is actually two things. It doesn't let you split tabs, which is pretty bad, of course. Second, it doesn't have full feature compatibility with my preferred coding agents. Another way I've seen this is that it doesn't have the slash goal command found in the Claude Code sessions. It only has it in Codex sessions.
This was a deal-breaker for me, so I had to stop using Conductor because of it.
Emdash
Emdash is the current program I use, 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 uses a terminal inside the application, and allows you to split windows.
The only downside, as far as I can see, is that Emdash doesn't have an equally nice arrangement of tabs on the left side where you see all your coding agents, that is, it doesn't have the backlog/in progress/update/organization you get in Conductor, but it still has a very nice way to organize the different agent tabs.
Term2
iTerm2 is also a good choice; it's just a basic terminal where you can run your own coding agents. However, I see no reason why; I stopped using the empty terminal for a reason: that while it works well because the coding agents are accessible via the CLI, there are no real features, and they don't organize the tabs that are convenient for you. So I stopped using Term2.
Claude code request
Claude Code app is very nice and very friendly. However, again, I don't like the way they organize the agent tabs on the left side and ongoing sessions. I just believe there are apps that organize different tabs better, like Emdash and Conductor.
It's worth noting that I use the Claude code program on my phone because I have a remote control. And this works very well, and I really like the application of Claude Code in this way.
Codex Application
Codex usage is basically the same as the Claude Code system, and I don't see any real difference for good or bad.
Omnara
Omnara is marketed as the kind of app that seamlessly syncs between your phone and your computer. While it synced well with the phone and PC, I found the overall app clunky to use, and I didn't find it very intuitive. It seems to me that the design was completely out of date, and that both Conductor and Emdash were simple, easy, and easy to use.
Cursor
Cursor was how I got started coding Agentic and tab completion in VS Code shortly after ChatGPT was released. The cursor is really good. However, if you only use Cursor, it is more expensive because most of the costs are based on usage and are not covered by subscription, like in Claude Code or Codex. And again, I don't believe I need to look at any code, so I see no reason to use a Cursor. Of course you can use pure agent view in Cursor, but at that time it is much better than other methods like Claude Code or Codex.
Another connection
Note that there are many other interfaces out there, and I haven't tested everything yet. It seems like there are new agent connections popping up every day. And it can be a lot of self-examination. What I urge you to do is to pay attention to the market for agent interfaces. If you see something interesting, quickly review their website, see if it fits your needs, and if it does, you should give it a try. I believe you will quickly notice if the code agent interface is worth using or not, say within twenty minutes.
How to find the right code interface
Now we need to get into how to find the right code interface. I've included a few things to consider and think about when choosing an agent interface:
- Does it have feature parity with all the new agent coding tools like Codex and Claude Code? If not, you will be behind. Which can be very frustrating when someone releases a feature that you want to try right away.
- Does it allow you to interact with your coding agents through your phone as well? I think this is very important. For example, if I use Claude Code, I can have something called remote control enabled, which allows me to get a session through my phone and have my coding agents continue to work and I can help them quickly when I'm not at my computer.
- Does it give you a nice overview of all your agent tabs and help you both remember which tab is doing what and which one has multiple sessions running at once without getting confused? Things that can help here are: automatic tab naming, separate tabs, workspaces, and more.
And that's it. All in all, I think you should, when testing the code agent interface, think about what is important to you and what is important to you. For some people, the most important thing is that it is very easy to use and accurate, while the rating of the feature is not so important. If so, I would say Conductor is a good choice. Some prefer a lightweight terminal above all else. If so, Warp or Term2.
In the end, you yourself have your own best taste preferences, and you should just try the things you find fun and see if they work better for you than the current interface you're using.
The conclusion
In this article, I have covered some methods of integration of the encoding agent. I started talking about why you should care about it and how it can affect how productive you are as an editor. I've covered many different links to coding agents that I've tried myself, covering the pros and cons. Ultimately, the interface of your coding agent is whatever you like to use the most. Different people have different preferences for what is important when dealing with coding agencies, and you should try different options yourself. I have compiled and listed quite a few examples for you to check out if you find them interesting. And I urge you to try a few different tools and spend some time on this because it can have a huge impact on how productive you are when working with your coding agents. It's worth investing some time to find the right coding setup for you.
👋 Touch
👉 My Free eBook and Webinar:
🚀 10x Your Engineering with LLMs (Free 3-Day Email Course)
📚 Get my free ebook Vision Language Models
💻 My webinar on Vision Language Models
👉 Find me on social media:
💌 Stack
🐦 X / Twitter
