What’s behind the OpenClaw ban wave
Abstract created by Good Solutions AI
In abstract:
- PCWorld reviews that Anthropic and Google are banning customers who join flat-rate Claude or Gemini accounts to OpenClaw attributable to extreme AI token consumption.
- Google DeepMind cites “malicious utilization” and repair degradation as causes for these unannounced bans, which supply no refunds or account reinstatements.
- OpenAI at present isn’t banning ChatGPT customers for OpenClaw utilization after not too long ago hiring the instrument’s creator, highlighting inconsistent business responses.
In the event you’ve been utilizing your flat-rate Claude or Gemini account to feed OpenClaw and its eye-popping AI talents, get able to be banned.
Particularly, those that’ve used their Claude and Google OAuth credentials for OpenClaw, the viral AI sensation that may gobble up hundreds of thousands of AI tokens in a single afternoon, have seen each Anthropic and Google ban their AI accounts. The bans ceaselessly come with out warning, and plenty of who’ve tried to search out out what occurred or get their accounts reinstated have been met with silence.
After all, you possibly can all the time dodge the banhammer by creating a brand new Claude or Google account, and it’s price noting that almost all banned Google AI customers are nonetheless capable of entry their Gmail, Google Drive, and different core Google companies.
Nonetheless, it’s not enjoyable to get locked out of the $200-a-month Claude or $250-a-month Google AI account you’ve been actively utilizing for OpenClaw and different instruments, together with Google’s well-liked Antigravity coding instrument, notably given the dearth of refunds.
“Fairly draconian from Google,” wrote OpenClaw creator Peter Steinberger on X, who added that he would seemingly take away help for utilizing Google’s Antigravity OAuth credentials to energy the viral AI agent. “Watch out on the market when you use Antigravity.”
Responding to the hubbub on X, Google DeepMind engineer Varun Mohan stated the corporate has “been seeing an enormous improve in malicious utilization of the Anitgravity backend that has tremendously degraded the standard of service for our customers. We would have liked to discover a path to shortly shut off entry to those customers that aren’t utilizing the product as meant. We perceive {that a} subset of those customers weren’t conscious that this was towards our ToS and can get a path for them to return again on however we have now restricted capability and wish to be honest to our precise customers.”
Whereas each Anthropic and Google are cracking down on using flat-rate OAuth credentials for OpenClaw, ChatGPT isn’t wielding the banhammer but–maybe as a result of OpenClaw creator Steiberger is now amongst OpenAI’s latest staff. OpenClaw itself remains to be an open-source instrument, albeit with the backing of OpenAI.
To grasp the hubbub over Anthropic and Google’s bans for OpenClaw use by way of OAuth credentials, you could know the distinction between OAuth and API entry to Claude and Gemini.
Even when you don’t know what OAuth entry is, you seemingly use it on a regular basis. Everytime you log right into a third-party service utilizing a type of “Login with Google,” “Login with Fb,” or “Login with Apple” buttons, that’s OAuth at work. Now, the Claude and Google bans for OAuth customers connecting to OpenClaw aren’t about OAuth per se; as an alternative, it’s the truth that OAuth credentials are used for authenticating flat-rate Claude and Gemini plans.
Usually talking, you’re not supposed to make use of your Claude or Google AI OAuth credentials to energy third-party AI instruments and companies, which usually don’t implement rate-limiting insurance policies; as an alternative, you’re supposed to make use of the Claude or Gemini API, which cost by the token reasonably than a flat month-to-month payment. (In AI lingo, a token is a tiny unit of knowledge utilized by AI fashions to course of your prompts and create content material resembling textual content, pictures, and video; the extra you work together with an AI, the extra tokens you burn.)
Nonetheless, simply since you’re not supposed to make use of your Claude, ChatGPT, or Google OAuth credentials for an unsanctioned third-party service doesn’t imply you possibly can’t.
For OpenClaw customers to authenticate their flat-rate Claude accounts with the bot, they will seize an OAuth token (to not be confused with an LLM token) by operating the Claude Code setup course of, then utilizing that token for OpenClaw reasonably than Claude Code. For Gemini customers, a Google Antigravity plug-in for OpenClaw does the trick. (The method for connecting ChatGPT to OpenClaw by way of OAuth is way simpler, however bear in mind, OpenAI simply employed OpenClaw’s creator and is thus extra invested in OpenClaw’s success.)
However whereas huge AI suppliers like Anthropic and Google might have tolerated such “hacky” OAuth authentication strategies up to now, they’re now cracking down in earnest, and the explosive reputation of OpenClaw has accelerated the method.
Not like earlier third-party AI instruments that roughly flew beneath the radar, OpenClaw went from obscurity to everywhere-all-at-once standing in a matter of weeks, far outstripping the person base of earlier instruments.
One other distinction is that OpenClaw’s jaw-dropping agentic AI talents come on the expense of tokens–loads of tokens, with some customers astonished to search out they’d used hundreds of thousands of tokens in a single afternoon. Merely asking your OpenClaw agent “how are you?” might chew by way of 30,000 tokens (or far more, you probably have a prolonged OpenClaw session going), versus a pair thousand tokens for ChatGPT on the internet.
From a enterprise perspective, Anthropic and Google most likely noticed cash happening the drain as OpenClaw customers let it rip with their OAuth-enabled flat-rate AI accounts. And to the purpose of Google DeepMind’s Varun Mohan, the OpenClaw-driven surge in OAuth utilization might have taken a toll on non-OpenClaw customers on the identical flat-rate plans. (Certainly, I seen frequent “making an attempt to achieve Gemini 3 Flash” warnings when vibe coding with Gemini CLI over the weekend. Had been these connectivity points have been attributable to OpenClaw utilization? Good query.)
To repeat, Anthropic and Google are more than pleased to fund your OpenClaw behavior by way of a pay-as-you-go API key, and certainly, some API-using OpenClaw customers have run up some eye-watering payments.
However so far as utilizing your flat-rate, OAuth-enabled Claude or Google account for OpenClaw, the get together seems to be over.

