Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Anthropic says OpenClaw-style Claude CLI usage is allowed again

docs.openclaw.ai

Post reply on HN
Print view View original

161–170 of 318 posts

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#161

Post by sumedh »

LatencyKills wrote: post #135
doginasuit wrote: post #103

I'm out of the loop on Claude, hasn't it always been possible to use the Anthropic API with a tool like OpenClaw, paying per request? Is this limitation just for using your monthly subscription account?

Many people likely objected to the original restriction because it seemed as though Anthropic was trying to impede the development of competing tools. If I'm paying for compute, why should it matter whether I use Anthropic's harness (e.g., Claude Code) or a 3rd-party harness?

Isn't their argument that third party harness dont play nice with their GPUs which is a fair argument.

With Claude Code they can predict what the traffic would look like with third party harness they cannot.

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#162

Post by pbronez »

JumpCrisscross wrote: post #152
Gigachad wrote: post #121

Earlier quoted context omitted.

Currently it costs so much more to host an open model than it costs to subscribe to a much better hosted model. Which suggests it’s being massively subsidised still.

If I drop $10k on a souped-up Mac Studio, can that run a competent open-source model for OpenClaw?

Rapid MLX team has done some interesting benchmarking that suggests Qwopus 27B is pretty solid. Their tool includes benchmarking features so you can evaluate your own setup.

They have a metric called Model-Harness Index:

MHI = 0.50 × ToolCalling + 0.30 × HumanEval + 0.20 × MMLU (scale 0-100)

https://github.com/raullenchai/Rapid-MLX

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#163

Post by JumpCrisscross »

pbronez wrote: post #162
JumpCrisscross wrote: post #152

Earlier quoted context omitted.

If I drop $10k on a souped-up Mac Studio, can that run a competent open-source model for OpenClaw?

Rapid MLX team has done some interesting benchmarking that suggests Qwopus 27B is pretty solid. Their tool includes benchmarking features so you can evaluate your own setup. They have a metric called Model-Harness Index: MHI = 0.50 × ToolCalling + 0.30 × HumanEval + 0.20 × MMLU (scale 0-100) https://github.com/raullenchai/Rapid-MLX

Pardon the silly question, but why do I need this tool versus running the model directly (and SSH’ing in when I’m away from home)?

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#164

Post by baq »

sofixa wrote: post #138
baq wrote: post #99

Earlier quoted context omitted.

> But the problem is, they have no moat I disagree very strongly with this, both anecdotally and in the data - subscriptions are growing in all frontier providers; anecdata is right here in HN when you look around almost everyone is talking about CC, codex is a distant second, and completely anecdotally I personally strictly prefer GPT 5.3+ models for backend work and Opus for frontend; Gemini reviews everything that…

You're actually reinforcing my point. Models are interchangable and easy to switch between to adjust based on needs and costs. That means that no individual model / model provider has any sort of serious moat. If tomorrow Kimi release a model better at something, you'd switch to it.

Yes, in that sense, technically correct.

I postulate in practice this won't matter since the space of use cases is so large if Kimi released the absolutely best model at everything they wouldn't be able to serve it (c.f. Mythos).

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#165

Post by LatencyKills »

sumedh wrote: post #161
LatencyKills wrote: post #135

Earlier quoted context omitted.

Many people likely objected to the original restriction because it seemed as though Anthropic was trying to impede the development of competing tools. If I'm paying for compute, why should it matter whether I use Anthropic's harness (e.g., Claude Code) or a 3rd-party harness?

Isn't their argument that third party harness dont play nice with their GPUs which is a fair argument. With Claude Code they can predict what the traffic would look like with third party harness they cannot.

If that was the argument, why did they reverse it?

Anthropic is constantly destroying goodwill and now seems to be in panic mode.

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#166

Post by giancarlostoro »

Uh, what? For the love of God can I make my own harness or not? Or is this just saying you can use it only in API mode?

I have had some ideas for a custom harness (like embedding some tools OOTB and replacing slow tooling) but these policies throw me off. Instead I use local models.

Problem is API costs are insane. I have toyed with the idea of running a local model that works with Claude Sonnet or even Haiku, and I know this has been done by others.

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#167

Post by rustyhancock »

zurfer wrote: post #107
rustyhancock wrote: post #75

Earlier quoted context omitted.

Consider Z.ai if you need "bulk" usage, GLM is now very good. They still have the occasional API brown out however. I used to use GLM mostly and had a Claude Pro subscription for occasional review and clean up. Now I just use GLM. I do think Claude Max is value for money. But it's more value than I personally need and I like Anthropic less and less.

Naive question but are you not afraid z.ai will train on your personal data?

I accept that all the providers will do what I would consider unethical with my data and simply don't expose what I don't consider a price of doing the business I want.

The other criticism I see is "ask it what happened in 1989" but as a my use case isn't writing a high school history essay I simply don't care. Or believe one should seek those kind of answers from any AI. (If you're curious it simply cuts off the reply).

I fully appreciate that YMMV and what sits right for others will not align with what's acceptable to me. Anthropic and OpenAI both are in my badbooks as much as Z.ai. pick your poison as they say.

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#168

Post by Atotalnoob »

JumpCrisscross wrote: post #152
Gigachad wrote: post #121

Earlier quoted context omitted.

Currently it costs so much more to host an open model than it costs to subscribe to a much better hosted model. Which suggests it’s being massively subsidised still.

If I drop $10k on a souped-up Mac Studio, can that run a competent open-source model for OpenClaw?

Qwen is probably your best bet…

Edit: I’d also consider waiting for WWDC, they are supposed to be launching the new Mac Studio, an even if you don’t get it, you might be able to snag older models for cheaper

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#169

Post by rcarmo »

PSA: Since you are still required to use Claude Code and I have had a bunch of non-technical people asking me to make https://github.com/rcarmo/piclaw based on Claude rather than pi (which is never gonna happen), I have started pivoting its Python grand-daddy into a Go-based web front-end that runs Claude as an ACP agent.

Still early days, but code is available, sort of works if you squint, and welcomes PRs: https://github.com/rcarmo/vibes/tree/go

Re: Anthropic says OpenClaw-style Claude CLI usage is allowed again

#170

Post by ElFitz »

stingraycharles wrote: post #59
1una wrote: post #18

Looks like this was restored 2 weeks ago[0], 3 days after Anthropic said OpenClaw requires extra usage[1]. At this point, it's hard to take this seriously. No official statement and not even a tweet? [0]: https://github.com/openclaw/openclaw/commit/d378a504ac17eab2... [1]: https://news.ycombinator.com/item?id=47633396

No, it's just that it's confusing, because there are two ways of using Claude Code credentials: 1. Take the oauth credentials and roll your own agent -- this is NOT allowed 2. Run your agentic application directly in Claude Code -- this IS allowed When OpenClaw says "Open-Claw style CLI usage", it means literally running OpenClaw in an official Claude Code session. Anthropic has no problems with this, this is complia…

And yet running the Claude Code cli with `-p` in ephemeral VMs gets me the "Third-party apps now draw from extra usage, not plan limits. We've added a credit to your organization to get you started. Ask your workspace admin to claim it and keep going." error.

One day you're experimenting just fine. The next, everything breaks.

And I'd gladly use their web containerized agents instead (it would pretty much be the same thing), but we happen to do Apple stuff. So unless we want to dive into relying on ever-changing unreliable toolchains that break every time Apple farts, we're stuck with macOS.

Post reply on HN