StartseiteArtikel

Another mobile app allows users to raise lobsters.

智东西2026-03-12 21:29
Get 50,000 for free, and they can be used for a month.

According to a report by Zhidongxi on March 12th, today, StepAI, an app under the brand of Jieyue Xingchen, has launched a cloud-based intelligent agent StepClaw built on OpenClaw. Users can deploy and call the capabilities of this kind of Agent with one click in the StepAI App.

In simple terms, StepClaw has moved the "lobster-raising" process that originally needed to be deployed on a local computer to the cloud. Users don't need to prepare a Mac Mini or a server, nor do they need to configure the environment by themselves. They just need to complete simple operations in the App, and the Agent can run continuously and execute tasks.

StepAI has opened 50,000 free trial quotas for StepClaw. After users download the StepAI App, they can claim a one-month trial quota, which includes 50 million Tokens for model calls, cloud server resources, and storage space.

Currently, StepClaw already supports deployment and use in the StepAI App, and the web version will also launch relevant functions on March 13th.

Click StepClaw in the upper left corner of the StepAI APP dialog box to enter the configuration page. (If you don't see the word StepClaw, you can try to restart the APP manually.)

01. No local deployment required, "lobster-raising" starts moving to the cloud

The so-called "lobster-raising" usually refers to deploying an Agent system on a local device, allowing it to run continuously and complete tasks such as information collection, content generation, or automatic operations. However, this process often requires preparing equipment, configuring the environment, and maintaining the operation. StepClaw's idea is to move this entire operating environment to the cloud.

Users enter the StepClaw configuration page in the StepAI App and click "Configure and use with one click". The system will automatically complete the OpenClaw environment deployment. The whole process usually takes only 1 to 3 minutes. At the same time, StepClaw also supports binding existing Openclaw.

StepClaw one-click configuration interface (left) and StepClaw chat interface (right)

After the deployment is completed, users will have a continuously running "lobster" in the cloud.

02. Runs 24/7 in the cloud, can continue working even when the phone is turned off

StepClaw runs in a cloud server environment, with the default configuration of a dual-core CPU, 4GB of memory, and 40GB of storage space.

Since the task execution takes place in the cloud, even if the user's phone is locked or the device is turned off, the Agent will continue to run. When the task is completed or there is new progress, the system will send a notification to the user.

The storage space is used to save historical information, enabling the Agent to gradually accumulate context and long-term memory.

At the model level, StepClaw is connected to Step's Step 3.5 Flash model by default. In addition, the system also integrates a skill tool library and Jieyue's self-developed search ability, enabling the Agent to perform real-time information retrieval and complete various tasks.

03. Conclusion: The threshold for "lobster-raising" is further lowered

In the past few weeks, Openclaw has gradually moved from the developer community to a wider user group. The complex deployment environment and high equipment costs are still the main thresholds that many users encounter when trying to "raise lobsters".

Cloud-based solutions like StepClaw attempt to further simplify the Agent operating environment. Users don't need to prepare equipment or build the system by themselves. They just need to complete the deployment in the App and can start trying to let the AI execute tasks continuously.

In the future, the way of using Agents may gradually change - from complex technical deployment to a more ready-to-use mode similar to ordinary applications.

This article is from the WeChat official account "Zhidongxi" (ID: zhidxcom). The author is Jiang Yu, and the editor is Xinyuan. It is published by 36Kr with authorization.