I fired it up, and the first thing I notice is that the arrow keys don't work. I can't select Claude Code. Oh, apparently it's in a different control, so I have to press Tab, and then the arrow keys work. Wow, this list of buttons has a slow scrolling animation when navigating it. Can I turn that off? Press enter on Claude, now I'm in a tiny modal window. Press enter, because I want to do the obvious thing, but apparently the obvious thing is "show in launcher", so the background of the modal is weirdly changing while a tiny single character inside the button is indicating that this is the part I'm supposed to be focusing on. No, I want to do the obvious thing of running Claude code. You could easily fit the 4 actions of this form on my screen, but by choosing to use a tiny modal window you're now forcing yourself to use another modal drop-down control to choose the action and a separate "yes actually do it" button, so the OBVIOUS ACTION of RUN THIS AGENT requires pressing tab, enter, down, down, down, enter, tab, enter. Great. Now I'm at a chat interface with an error screen, because it isn't installed. Quit the program, restart, enter, tab, enter, down, down, enter, tab, enter to install. It shows a successful run of the "ACP adapter" for claude. Shift-tab, enter, down, enter, tab, enter. Now I'm back at exactly the same error screen because apparently the install didn't work. Now, I know that you need to be running "npx @zed-industries/claude-code-acp", so I check the docs and apparently I can "toad run COMMAND". But it doesn't work for multi-word commands. And my trial with toad comes to an end.
So I can't test it for anything actually useful right now, but I'll add this to my list of projects to watch. Hopefully, being a UX-focused project, the creator actually focuses on the UX and fixes some of these silly decisions.
It's possible this is just part of the learning curve, but it is making me think I'll have to come back to this project in a month or two to see if there are fewer pain points. Great work so far though.
Textual offers incredibly impressive terminal experiences so I'm very much looking forward to this.
I wonder how much agentic magic it'll be able to include though - Claude Code often seems like a lot of its intelligence comes from the scaffolding, not just the LLM. I'm excited to see!
First, from my reading of the ACP doc, one thing that seems pretty janky is if the ACP client wants to expose a tool to the agent, e.g. if Toad wanted to add the ability for the agent to display pretty diffs. In the doc they recommend stdio to the ACP server, then stdio to an MCP server, and then some out of band network request back to the ACP client. Have you thought about this, or found a better solution working on Toad?
Similarly, it would be useful to be able to expose a tool which runs a subagent using ACP using a different agent, e.g. if I'm using Claude for coding but I'd like to invoke codex for code review. Have you thought about doing anything like this? Is it feasible over the protocol?
One of the advantages of Toad is that it is vendor agnostic. In the future Toad will be able to run sub agents, and allocate any agent to any job. Still to figure out the UX for that.
1. How has it been working with ACP? Is it anywhere near feature parity with Claude code’s native interface?
2. I see your repo is written in Python which is interesting to me for a responsive TUI. Is it snappy and performant and if so what gave you done to make it feel native? And why did you choose Python?
Python is more than capable of running a TUI. It is just text manipulation after all. Toad uses Textual, which is currently the best TUI library around. I may be biased saying that as I built it...
I was about to try opencode after using claude code for quite a while.
I think understand the fundamental difference in how they work (acp against existing agentic loops with toad vs a single agentic loop for all models with opencode) but I’m curious why we might want toad over something like opencode, which lets me use any model under the sun.
I suppose toad gets to use the highly specialized agentic loops for each cli. And has a nicer (? opencode is pretty slick from my brief usage…).
Curious to hear about why you chose to built this way and what advantages you see.
You get a nice UI that is only going to get better as time goes on.
It’s far better model to separate the agent from the UI. The current situation is like building a browser for a single website.
How are new agents added? Do you have to write a dedicated plugin for each one? Or there's some kind of discovery mechanism?
(I was looking for Copilot, but I guess that will depend on https://github.com/github/copilot-cli/issues/222 ?)
I think they are working in the Copilot ACP layer. Doubt it will take long.
And Copilot isn't supported for now because, well, there is no ACP support
Ideally working with toad to experiment with it.
If an agent can be configured to use Ollama, then you could use it from a Toad. It might be possible right now.
Check out vibecommander - it’s a young tool in this space with a different take that wraps around CLI coding assistants with IDE-style file and git panels that compliment the experience by letting the human do the code review part of the task seamlessly.
Will add Toad support ASAP, I’m sure they’ll be great together.
Stop trying to convince people to use vim, it sucks, it's got a terrible ux, it's not intuitive, it's overly complicated, hard to learn, arcane, and looks like ass.
It literally is using Claude under the hood. Should be no different than Claude’s own CLI.
Reading the article I felt like the "sickos: yes!" meme.
It worked perfectly and looked splendid doing so.
Excited to dig in further.
