Just now, Claude has completely lifted the ban on third-party applications.

Claude is making a triumphant return.

Just now, Anthropic dropped a game-changing move, and Claude is about to see explosive growth!

In short: Anthropic has completely removed policy barriers, officially confirming that the quota for Claude subscription users can be seamlessly accessed by third-party Apps and Agents!

This is a key step for Anthropic in the second half of the large model competition.

From now on, the "double charging" curse between developers and users has been officially broken.

The great navigation era of the Agent ecosystem will be fully launched!

The Sword of Damocles hanging over developers' heads has broken

Claude's platform policies often change overnight, which plagues the vast number of developers a lot.

In the past few months, a worrying interpretation has been circulating in the community: Anthropic may restrict or even block third-party applications from calling the Claude API and Agent SDK.

The reason is very simple — no benefit should flow to outsiders.

Large AI manufacturers also want to build their own ecosystems, why should they let third-party developers make money with their underlying models?

This concern is exactly the Sword of Damocles hanging over the heads of developers. No one dares to bet their entire career on it.

But now, Anthropic has completely eliminated this concern.

On October 7, Anthropic quietly updated its official support page — from now on, all third-party applications are fully approved!

Link: https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan

The document clearly states that for users of the Claude Max and Team plans, their monthly API Credits can be directly used for:

Claude Agent SDK

Claude -p Command

Claude API

Claude Managed Agents

All third-party Apps that use the above technologies

As soon as this news came out, netizens immediately celebrated and shared the good news with each other ——

From now on, third-party applications that use the Claude Agent SDK are allowed to consume the subscription quota!

To completely dispel doubts, many official Anthropic employees even personally showed up in the comment section of Hacker News (HN) to confirm this policy by posting replies.

"There are no objections at all."

"We support the continued use of this technology."

"No extra fees are required."

Every word is firm and unambiguous.

The policy is now clear. The red line is gone. The alert is lifted.

Why is this a historic deregulation?

To understand the depth of this carnival, we must understand the past AI application ecosystem.

In the era of large models, the business model of applications has always been a huge pain point, which can be called the "double charging paradox".

As a user, you spend hundreds of yuan a month subscribing to ChatGPT Plus or Claude Pro. As a developer, you have made a great programming assistant App that accesses Claude's API at the underlying layer.

When users use your App, the API will generate costs. Anthropic will charge you (the developer) for this money.

Then, how can you (the developer) survive?

You can only charge users.

Then, an absurd scene appears: Users have already paid for Claude's membership, but they have to pay the developer a second time when using third-party Claude applications.

This is double payment.

This directly leads to two fatal consequences:

First, users are unwilling to pay.

I already have a Claude membership, why should I pay you?

Second, developers cannot act boldly.

The API call cost is too high, and I will go bankrupt after running a few complex Agent tasks.

This forms a dead end. A large number of excellent creative ideas die before they are implemented. The sunk cost is too high, and the fault tolerance rate is too low.

Anthropic's new policy, which fully lifts restrictions this time, essentially establishes the "BYOS" (Bring Your Own Subscription) model.

You developed a third-party App? No problem.

Users want to use it? No problem either.

As long as users are members of the Claude Max or Team plans, the computing power they consume in your App will be deducted directly from their own monthly API quota!

What does this mean?

For users: One pass, free access across the entire network.

After purchasing a premium Claude subscription, you can not only chat on the official website, but also seamlessly connect to various excellent third-party automation tools, code completion plugins, and data analysis Agents. No secondary payment is required, and there are no additional thresholds.

Your subscription has truly become a "pass for the AI era".

For developers: Zero cost, focus on innovation.

You no longer have to stare at the API bill in fear! You can boldly design the most complex Agent architecture and call the most computing-intensive Chain of Thought.

As long as users bring their own quota, your computing cost is zero. You can put 100% of your energy into product experience and functional innovation.

Claude developers can now use their Claude subscriptions in third-party environments, no extra fees required!

This not only saves money, but may also completely liberate productivity.

Anthropic's open conspiracy: Crush competitors with "openness" to win the second half of the competition

Anthropic made this decision by no means a sudden change of heart.

This is a commercial open conspiracy.

The competition of large models has entered the second half.

In the first half, everyone competed on the number of parameters, benchmark scores, and the "chat" intelligence.

In the second half, everyone competes on action capability and Agent ecosystem.

Whoever can make AI access real business flows, whoever can make AI automatically write code, send emails and operate software, will win the future.

At this critical juncture, the routes have diverged.

Some giants have chosen the "closed empire" model.

They try to enclose all traffic, all applications, and all Agents in their own Apps. Want to use my model? Go to my app store. They act as both referee and player, trying to squeeze every drop of profit in the industrial chain.

But Anthropic has chosen another model.

Anthropic is very clear that the power of one company is limited. It is impossible to develop the best Apps in all vertical fields around the world by itself.

Rather than guarding developers as potential competitors, it is better to turn them into their own "workers".

Allowing subscription quotas to directly "freeload" third-party Agents is Anthropic's killer move to accelerate the prosperity of its own application ecosystem.

Just imagine the following scenario:

Without the worry of API costs, applications developed based on the Claude Agent SDK will see explosive growth.

Today, an Agent that automatically edits videos comes out.

Tomorrow, an Agent that takes over the reconstruction of your entire codebase comes out.

The day after tomorrow, an Agent that automatically helps you process massive Excel spreadsheets comes out.

The better these third-party applications are, the more people will use them.

And if users want to use these god-level Apps, they must purchase subscriptions for Claude Max and Team!

Anthropic is using the wisdom of the entire third-party developer community to help itself sell memberships!

By lowering the usage threshold of third-party applications and the sunk cost of developers, it cleverly deeply binds the interests of global developers to its own.

Developers get free computing power channels and a large number of high-value users. Users get the most prosperous and hardcore tool ecosystem on the market. Anthropic gets a steady stream of subscription fees and an unshakable ecological moat.

This is an open conspiracy. Competitors can only watch Anthropic's ecosystem grow like a rolling snowball, but it is difficult to replicate this strategy — because it requires extremely high confidence and full trust in its own model, especially the code and Agent capabilities.

Time to act: Developers, what are you waiting for?

The window of opportunity is always fleeting.

Looking back at the history of the Internet, every time the underlying policy of a platform is opened, a number of great companies will be spawned.

Apple opened the App Store, which led to ten years of prosperity for the mobile Internet.

WeChat opened mini-programs, which supported countless startup teams.

Today, Anthropic's clear support for third-party subscription calls is the "App Store opening moment" in the AI large model era.

For the vast number of independent developers, product managers and startup teams, this is a once-in-a-lifetime dividend period.

Don't hesitate any more.

Read the documentation. Write code. Call the SDK.

Use the sharpest Claude model to cut through those outdated, inefficient and painful business scenarios.

Create truly implementable Agents.

The golden age for action takers officially starts today.

We have only just begun.

References:

https://x.com/daniel_mac8/status/2109686606647181653?s=20

This article is from the WeChat official account "Xinzhiyuan" (ID: AI_era), written by David Aeneas, published with authorization from 36Kr.