It works off the Claude Code SDK, which mean it doesn't support many of the built in slash commands - it doesn't support /compact, which is 100% necessary because when you use this implementation enough, you'll eventually get a "Prompt too long" error message with no ability to do anything about it. Since you can't see how far you are in the context window, it's a deal breaker, since you have to start a fresh chat and might run out of room before you can ask it to create a summary prompt for continuing.

There is no way to switch models that I can tell - I think it just picks up on your default model - and there is no way to switch to Plan mode, which has become absolutely crucial to my workflow.

I didn't see Zed picking up on problems reported in the IDE, it was defaulting to running 'tsc -b' in my directories.

At this point it's better to run a terminal inside Zed and work from there. The official response in the Zed Discord has been "talk to your local Anthropic rep" to get them to support Zed's Agent Client Protocol (ACP).

One thing that still suffers is AI autocomplete. While I tried Zed's own solution and supermaven (now part of Cursor), I still find Cursor's AI autocomplete and predictions much more accurate (even pulling up a file via search is more accurate in Cursor).

I am glad to hear that Zed got a round of funding. https://zed.dev/blog/sequoia-backs-zed This will go a long way to creating real competition to Cursor in the form of a quality IDE not built on VSCode

My difficulty in finding editors that fit my desired input scheme kinda reminds me of the old pre-LSP days. Where you'd chose an editor based on it's language features. I wonder if we need some sort of common editor interface to allow these sort of text editing primitives to work in new editors, as it seems to be considerable friction.

But, I find Zed challenging to adopt due to random nuances. First, settings management is a mixed bag and sometimes I just want a quick way to open the "settings.json" from the settings pane without fussing around. Then I'd like the "settings.json" to stay open (reopen) on a restart of Zed. Then I'd like the ability to use an LLM that doesn't have native tool calling support, which Zed seems to be the only app I've used that doesn't have a workaround. Then I'd like the UI to be a little easier to navigate as a new user, it feels a bit scattered and overwhelming at times.

I haven't used Zed much and I may give it another shot (soon), but it very much feels like a tool built by engineers for engineers... Which is great for power users, but seems not so great for new adopters.

I don't think the shortcomings are a blocker, but they are the reason I haven't adopted Zed. The shortcomings are just enough for me to take a step back and say "maybe I'll try again later".

- I don't want to constantly auto-accept. The point of auto-accept is that it auto-accepts. Seems like a bug.

- It'd be great if I could go back to a specific message and delete the ones I don't want, similar to the CLI version.

- Where is Plan Mode? Maybe I just couldn't figure out how to get to it.

- I can't easily see Background Tasks.

- How do I change models?

- How do I create new sessions (via /new for instance)? Why is `/clear` not supported?

- I don't want to see the entirety of the edits in the terminal. Can they be collapsed by default? Or maybe show a preview?

I just hope I'm wrong about the medium term impact of the VC funding but rushing AI AI AI out seems to be a sign of that rather than fixing fundamental issues that remain such as the ugly font rendering.

I have a $200/month Anthropic Max subscription that I use for help in exploring and coding my math research. As of now no AI model can compete with Opus 4.1 for helping me with my most challenging tasks. I try every one I can. Gemini 2.5 Pro is great for code review and a second opinion, but drives off the road when it takes the wheel.

I tried a $100/monthly plan and spent $20 in an hour the first time I went over; an API key is not a practical way to use Opus 4.1.

There are plenty of concerns using Clause Code in a terminal, that Zed could address. Mainly, I can't "see over AI's shoulder" so I need to also test. The most careful extension I coded was terminal sessions we could share as equal participants. Nevertheless, as a rule I'd attribute my relative success to just living with shortcomings, as if a "partner that snores". AI loses track of the current directory all the time, or forgets my variable naming and comment conventions? Just keep going, fix it later.

How can I get equivalent value to my Max plan, using Claude Code Opus 4.1 with Zed?

https://x.com/sridca/status/1963271904384401886

There are a lot of comments that people need X feature in order to switch to Y editor. While that may be true and your particular workflow requires certain features, what is overlooked is the survival pressure for editors.

It appears that our industry is moving towards adoption, sometimes mandatory, of AI coding agents. Regardless of your feelings on the topic, having good tooling to support this effort comes down to: switching costs, compatibility with existing editors, and a strong ecosystem of third party extensions.

While Cursor/Windsurf jumped the gun on bespoke editor integrations with LLMs - the adoption of MCP and other SDKs for coding agents means it's plug and play. The full feature set will be in every editor connected to every agent.

I think Zed wins on having the lowest switching costs for most developers. Paying down generic solutions like Agent Client Protocol (AC) now is a good strategy. It took multiple parties coming together for us to get TLS, OAuth 2.0, and ECMAScript.

I don't see why most editors should behave like hand crafted musical instruments when in reality they are much more akin to high quality knives in a kitchen (sure you have your favorite knife set and bring it from job to job, but at the end of the day you can be just as productive with a different knife when necessary).

So if Zed automatically handles that (where there's a worktree per thread) I can see the appeal. Apart from that, I'm already using Tower to view the changes so I'm not really sure what the value here is.

I tried installing it, and got an error "can't load supported slash commands" – not sure what that means.

As someone who’s running a development agency I need to have tens of dev environments for different client projects running at the same time, and being able to switch between them multiple times every day (often from multiple client computers), so a remote server is the only way to go–I don’t want all of that stuff running on my Macs.

Nowadays I also have tens of CCs running on the dev server, switching between them using tmux, which works great, but the lack of support for pasting images through the terminal/ssh/tmux has been a real bummer. It would be great if Zed found a way to bridge that gap.

But tbh if it’s not as frictionless as the Codex IDE extension in a Zed-skinned VSCode, it doesn’t matter.

Tried giving Claude and CC many chances, but the cognitive load of constantly managing a hard context window is DOA.

Codex w/gpt-5 is on par if not better than any of Anthropic’s solutions at this point, and the ubiquity (web, CLI, IDE) + UX consistency of Codex under one account/plan just dominates any marginal value of using a different model at a higher price.

Codex just works. Then it keeps working. Then it keeps working.

Any solution that wants to compete with OAI’s latest hostile takeover attempt has to match then beat on “unlimited/anywhere/frictionless” UX across platforms AND price ($200/mo all in).

I don’t see a good way out of this for most, except through major spend on playing catchup.

Guess that’s why Anthropic just raised again. Cursor is clearly trying to play, but they will always be a markup product until they launch their own SOTA model. Is Gemini still alive?

How do I configure it with Base API URL?

[0]: https://news.ycombinator.com/item?id=45045515

What if I want to send a subset of my open editor panes to Claude Code? What if I want Claude Code to open diffs for its edited files in a specific area/window, and silently open that file so that I can multitask on other things without it taking focus when it's done thinking? What if I want keyboard shortcuts for specific slash commands, or to trigger a slash command from another task?

Having a robust open-source ecosystem that will let users fork and build customizable UI around coding agent experiences will make them even better, and the space will move even more quickly because the ecosystem won't split between different preferences for agent/model choice. It's an incredible time to be coding.

Zed doesn’t have file delete undo and I found the AI autocomplete to be so bad I had to turn it off. I still use vscode for work. Zed is kind of in a purgatory state where its lacking in way to be a main for many folks but its close. Just the AI focus is sadly the best way for it to keep investments going, but are really not the top things I think Zed needs.

Zed has a lot of these micro-battles ahead where it has to spend money building solutions that VSCode's community shipped without their core team putting in any effort at all.

Agent Client Protocol (ACP) - https://news.ycombinator.com/item?id=45074147 - Aug 2025 (93 comments)

I've been seeing a lot of crashes of late

I can't believe people are ok with horizontally layed out tabs.

Internal error: { "details": "can't load supported slash commands" }

That wasn't stated or perhaps I missed it.

> Escape the Terminal

it does not sound like a good thing to me: a) the terminal is fine; b) AI should not escape anything.

I have a subscription to Claude. What gives?

Zed's dead baby, Zed's dead

I keep trying this editor every few months ever since it was announced, but I always have similar experiences. I remember once I managed to start actually editing something before the GUI started to disappear.

But hey, it has AI.

Yesterday I looked again at LunarVim's website. While LunarVim seems to look pretty, it has a lot of dependencies, including pip, npm, and more. Seems like it is installing stuff from everywhere. Not so confidence inspiring, especially pip and npm installs.

And just now I see this Zed blog post linked on HN! But, unfortunately the website is not inspiring much confidence either. Can anyone explain to me, why I cannot see any _text_ on all of zed.dev, without running JS? I mean, I probably know the answer, or some possible answers, but man, that's already such a turnoff, I already doubt the editor is any good now. Would be good, if they could fix their website, and make simple text, simple text again, accessible and all that. Please get some craftsmanship into this website.

EDIT: 'pparently I said something some people don't want to hear, lol.