Anthropic just BANNED OpenClaw usage... What Now and What you should switch to.

If you woke up to an email from Anthropic and your stomach dropped - yeah. Same.

Starting today (April 4th, 8pm BST / 12pm PT), your Claude Max subscription no longer covers OpenClaw usage. Or any third-party agent tool, for that matter.

That $200/month you've been paying? It now only works with Claude's own apps. Claude.ai, Claude Code, the mobile app. That's it.

Your OpenClaw setup? Your agents? Your cron jobs?
Your entire workflow you spent weeks building?

They want you to pay again for all of that. On top of your subscription.

This is probably the outcome we've all been waiting for - and dreading.

Anthropic announced - with roughly 24 hours notice - that Claude subscriptions will no longer cover "third-party harnesses." That's their word for OpenClaw, and basically every agent framework that isn't made by Anthropic.

Your existing setup still technically works. But after the cutoff, every message your OpenClaw agents send gets billed as "Extra Usage" - pay-as-you-go, billed separately from your subscription.

So you're paying $200/month for Claude Max. Plus per-token costs for every single thing your agents do through OpenClaw.

Double dipping. No other way to describe it.

You've got options.
Here's what they actually look like.

Option 1: Stay On Max + Pay Extra Usage

Keep your $200/month subscription.
Keep using Claude through OpenClaw.
Accept the extra billing.

Who this is for.
People who don't want to change anything right now and are happy paying more for the convenience. Without changing to a full API.

The reality.
You're probably looking at $300-500/month total depending on how many agents you run and how active they are. Heavy users will pay more. The Max subscription gives you unlimited Claude.ai usage still - so if you use Claude directly a lot AND run OpenClaw, this might make sense.

The problem.
You're paying a premium for loyalty to a company that just pulled the rug. And the per-token costs are identical to Option 2 anyway.

Option 2: Switch To The Anthropic API

Ditch the subscription entirely. Or keep it for Claude.ai personal use and run OpenClaw on a separate API key.

Go to console.anthropic.com create an API key. Load up some credits. Tell your OpenClaw to use that key instead. Back to the start, what we all ran in the first month of using OpenClaw.

Pricing (per million tokens):

Who this is for.
Anyone who wants to keep Claude's quality but control their costs. You set your own spend limits. No subscription required.

The reality.
Most people running a couple of agents will spend $200-400/month on API tokens. Roughly the same as Max was costing you anyway. But now you can see exactly where every dollar goes. And if you're smart about using Sonnet for grunt work and only Opus for the thinking - you can keep it under $200 easy.

What to do right now.
Even if you're not sure which option to pick long-term, load $50 onto the API and switch your OpenClaw over today. Gives you a smooth transition while you figure out the rest. You have until 8pm BST before they force your hand anyway. Or better yet - just use the 'free credit' they are offering subscription users right now as your transitional credits.

Option 3: Switch Models Entirely

Here's where it gets interesting.

You don't have to use Claude at all.
OpenClaw supports multiple providers. And honestly, the competition has caught up.

OpenAI + Codex.
The obvious play. OpenAI literally owns OpenClaw now (they acquired the creator). GPT models work natively. Codex is built for coding agents. The integration is only going to get tighter. If Anthropic is pushing people away, OpenAI is rolling out the red carpet.

Google Gemini.
Massive context windows. Competitive pricing. Good for research and analysis heavy agents.

Kimi.
Surprisingly good. Cheap. Worth testing for specific use cases.

Nous.
Open source route. Run locally if you've got the hardware.

Who this is for.
Anyone who's tired of being at Anthropic's mercy. Or anyone who wants to cut costs dramatically.

The reality.
Claude still has the best "personality" for agent work. The way it follows instructions, maintains context, handles complex tasks - it's ahead. But GPT is close. And if OpenAI puts real effort into making their models work better with OpenClaw (which they absolutely will now), Claude's advantage shrinks fast.

Option 4: Use A Hosted Service

Services like Hermes and others will host your OpenClaw setup and handle the model routing for you. They negotiate bulk rates, manage the billing, and you pay one monthly fee.

Who this is for.
People who don't want to think about API keys and token counts. Just want their agents to work.

The reality.
You're adding a middleman. Could be fine. Could add latency. Depends on the service.

My Recommendation.
I actually have an article scheduled this week covering a hosting service I really like, you'll have to stay tuned for it - but I think this could be a great switcharoo for anyone trying to make this whole thing easier.

I'm not going to tell you what to pick.
I'm going to test it all and show you.

I run 6 agents across two servers. Multiple skills, cron jobs, a full dashboard, the works. This change hits me hard. Especially because I just spent 48 hours recreating my entire setup and building a really fucking cool 'space' dashboard... smh.

But here's my plan. I'm switching Momo (my main agent) to the Claude API. Full Opus. Because I need the quality there and I'm willing to pay for it.

Rocky (my builder) is moving to Codex. He writes code all day. That's what Codex is built for. Lets see how crazy the switch is.

Dr. Grace (my scanner/ops agent) stays on Claude Sonnet through the API. Sonnet is cheap and he does volume work.

Mary (my creative tool & writer) will also move to an OpenAI GPT model and test how well it fares in comparison to Opus. This might be the only other model that needs to remain Opus but lets find out.

Fergie and Gaffer (specialist agents) are moving to OpenAI GPT. Good enough for their use cases and way cheaper.

I'll run them side by side over the next week. Same tasks. Different models. And I'll publish exactly what happens - the good, the bad, the "holy shit Claude is actually better" moments and the "wait, GPT handles this fine" surprises.

Anthropic just made a business decision. They looked at how many people were running entire agent operations on a $200 subscription and said "nah."

Fair? Depends who you ask.

Smart? Probably not.

Because here's what they've actually done. They've given OpenAI - who owns OpenClaw - the biggest gift imaginable. Every OpenClaw user is now actively shopping for alternatives to Claude. And OpenAI is standing right there with open arms.

Claude might still be the best model for agent work today. But Anthropic just gave everyone a reason to find out if that's still true tomorrow.

1. Before 8pm BST - switch your OpenClaw to API key auth or decide to accept Extra Usage billing
2. Load $50 minimum onto your Anthropic API account as a safety net
3. Decide which agents actually need Opus - most of them probably don't
4. Test one agent on GPT or Codex - pick your least critical agent and switch it. See how it goes
5. Follow along as I publish my model comparison results over the next week

The sky isn't falling. Your agents aren't dead. You just need to make a couple of decisions today instead of next month.

And honestly? This might end up being the push that makes everyone's setup better. Cheaper. More resilient. Less dependent on one company's pricing whims.

That's the silver lining. If you choose to take it.

If this helped, share it with someone running OpenClaw who hasn't heard the news yet. They've got until 8pm.

For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.

LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.

The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.

Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

Learn more.