I mostly get this change from their side except for the extremely misleading framing. I think in response to this change, I might write an open-source Claude Code GUI wrapper to wrap over the TUI and allow accessing it from a GUI interface.
Still interactive (conforms to spirit of Anthropic's new rule), and runs Claude Code via the actual CLI (conforms to letter of Anthropic's rule), without even using the JSON output flag. It would just read the outputs directly from the terminal like a user would, and then translates the user's actions directly into actions to the Claude Code TUI.
Tradeoffs: the thinking blocks are lost, and file changes from edit tools might have to be reconstructed after-the-fact using VCS commands because the TUI doesn't display them when accept-edits mode is enabled. Requires underlying terminal to have a giant width/height in order to reliably capture the outputs by reading the terminal alone. Interactive tools like AskUserQuestion will need to work via detecting that such a prompt is open and exposing the original CLI interface to the user temporarily.
But overall it seems like it could give users most of the benefits of an interactive GUI interface (scrollable history, much higher info density, easy to copy text, browse features with mouse) while staying compliant with Anthropic's new rules both in letter and spirit. You can then build a worktree manager or whatever above this agent GUI layer. This way us interactive users can have our good tools back (and leverage the new claude -p credit for occasional automation features) while the people who are actually abusing the plan with 24/7 dark factories or openclaw are still locked out.
(the only way Anthropic would not agree to this is if they had the ulterior motive of lock-in in mind...)
Sounds like you're thinking of scraping the raw terminal output. Consider using a terminal library (pyte[0] comes to mind) for interaction and processing the session files for tool use data, etc.
Yeah i've been thinking of hooks and reading the session file as possible ways for getting tool use data.
Also I managed to get a VSCode Claude Code wrapper working last night. It launches instances of browser-based VSCode (via Coder's code-server) with Claude Code plugin pre-installed, then hides all toolbars and opens the CC window using another bundled custom plugin. Still have a few integration problems to solve, but in theory this should be perfectly allowed by Anthropic, give you a pretty good UI, and still allow integration into an agent orchestrator using hooks/session files to get the agent data.
It's a bit disheartening that this is generating so little discussion (this thread seems to be the one with the most comments, currently at 8).
99% of my max plan usage is non-interactive, and this post-June 15 pricing will far, far exceed what I can afford. I assume this applies to a great deal of us.
is using a Ralph loop for all tasks with "claude -p" and using my weekly limit up to 100% considered some kind of unprofitable outlier? It's a command line tool, it would be ridiculous to expect that a large number of users don't do this. I never launch "claude" interactively by itself.
My understanding is that now with this workflow I will pay the same amount but get much less usage before getting to 100% for the week. How is that not bad?
They've probably (rightfully) identified long horizon autonomous development isn't quite there yet, so most stuff created by forcing Claude to run until you hit 100% of your weekly limits is going to be relatively low value slop that will not convert to long-term sticky spend/usage vs someone actively steering Claude until they hit their limit.
If anything you might be an outlier if you do this and are actually producing something of reasonable value down the line.
They know if the value provided by subsidized usage is too low, they're literally burning dollars for nothing: people won't feel any great pain losing it one day and definitely won't convert.
Lots and lots of automation, along with long-running multi-instance headless tasks. My usage, which I was on the verge of increasing via a second Max plan, would cost 3 figures per day, rather than per month. I already know this because my initial tooling was prototyped against direct API usage, and that was the cost of my usage patterns.
That is not sustainable for me. I understand that Anthropic can't subsidize all of us forever, but this will force me away from claude which is unfortunate. A number of the people I work with are similarly affected.
Same. Lots of little tasks that add up over time. The economics might not be great for Anthropic, but the economics of this change definitely don’t work for me; I could build a home lab and recoup its cost in about three months for the same pricing. “Headless is expensive” is a weird tone to set given so many recent cc releases pushing for greater autonomy (agents view, auto mode).
I’ve been happily single-homed on Anthropic for the last year, but I’ll be spending the next month rewriting projects off it to take advantage of better value elsewhere. I suspect that’s a one way door.
I'm curious what you're doing so often and repeatedly that can't be done with more normal software. Or maybe write a good school, I've noticed Haiku+good instructions often matches Opus
This could end up becoming a cat-and-mouse game where users programmatically try to turn their non-interactive usage of Claude Code to appear interactive and Anthropic tries to detect and charge that under API pricing. I don't know if there's a proper solution here because there will always be borderline use cases like using Claude Code on a cloud VM, where it would be nicer to interactively do work through sending and receiving messages on a custom frontend rather than SSHing and using the CLI.
Next thing to go will be loops and scheduled tasks, if they keep needing to trim usage to fit in the available compute, I suspect, if ‘claude -p’ is essentially gone, now.
I wasn't able to post the official links because it was already posted by simple reposters who didn't understand, and titled their HN posts as "you get credits" basically
Ya but that monthly usage doesn't refill every five hours or every week the way your subscription usage does. Once you're through the $200, you're done until next month.
These Twitter/X links are killing me, especially when they don't work with XCancel which seems to be happening more frequently. This one just gave me a summary by Grok, and I have no idea now if I'm seeing the same thing that others are commenting on.
Happy to change the title if there is explicit clarity on what my `claude -p` usage costs me today & it not differing from what it'll cost me once the change is in effect.
I wonder how this cat-and-mouse game will end. I've been using my own wrapper around Claude CLI that overcomes some 'claude -p' limitations, and now I have even more reasons to.
But inevitably, this will end, ultimately making AI a bit less accessible for consumers.
As a fellow zed user, I fear this could affect us (though i'm not 100% confident). Zed uses their own adapter to implement claude's ACP support, and it's likely using `claude -p` (the agent SDK) under the hood.
It's really important to note that the programmatic vs. interactive framing is a little misleading. For example, if you're using some IDE integrations, that is still interactive, that is still the same usage pattern as using it in the Claude Code CLI. BUT with these new rules, it'll count against this special "programmatic" usage.
Considering what kind of fuckups happened up until now everytime Anthropic vibe codes a horrible anti-automation "fix" (like billing extra usage just because the string "hermes" exists somewhere in the prompt), this will get funny/horrible really fast. After all, you can very easily script tmux to use Claude Code "interactively".
What crazy shenanigans are gonna happen? Are people that type too fast, or people that copy paste too much, or people that output to markdown/json too often going to wake up to a 800$ extra usage bill or a banned account?
This is much of a something burger for users who actively use `claude -p` under their subscription. Users will have to do their own math, but that 200 could come and go quite fast and then you're hitting normal API rates versus what you had previously been hitting.
Sure, but if you spend that many tokens you're probably a pretty big net loss for the company anyways. No company will subsidize that level of compute for you indefinitely.
Still interactive (conforms to spirit of Anthropic's new rule), and runs Claude Code via the actual CLI (conforms to letter of Anthropic's rule), without even using the JSON output flag. It would just read the outputs directly from the terminal like a user would, and then translates the user's actions directly into actions to the Claude Code TUI.
Tradeoffs: the thinking blocks are lost, and file changes from edit tools might have to be reconstructed after-the-fact using VCS commands because the TUI doesn't display them when accept-edits mode is enabled. Requires underlying terminal to have a giant width/height in order to reliably capture the outputs by reading the terminal alone. Interactive tools like AskUserQuestion will need to work via detecting that such a prompt is open and exposing the original CLI interface to the user temporarily.
But overall it seems like it could give users most of the benefits of an interactive GUI interface (scrollable history, much higher info density, easy to copy text, browse features with mouse) while staying compliant with Anthropic's new rules both in letter and spirit. You can then build a worktree manager or whatever above this agent GUI layer. This way us interactive users can have our good tools back (and leverage the new claude -p credit for occasional automation features) while the people who are actually abusing the plan with 24/7 dark factories or openclaw are still locked out.
(the only way Anthropic would not agree to this is if they had the ulterior motive of lock-in in mind...)