Since its launch, OpenClaw quickly caused global fever. The AI model runs right on the machine, has long-term memory, performs tasks automatically. Many practical applications of this software are shared, causing a wave of widespread interest. the world.
However, underneath the description of Free Forever was originally a huge resource system to run the tool. Users were forced to invest large costs for powerful machines or continue paying monthly API leases from leading companies.
It's free, it costs money.
In Vietnam, when you search for the keyword "OpenClaw Installation Guide", there are hundreds of The video.With popular operating systems such as Windows, MacOS, and Linux, the setup also requires copying a few lines of code to run in PowerShell or Terminal. However, the lack of user interface may present some difficulties for individuals with little programming experience.
Interchange with Knowledge - Znews, Luân Ngô, an AI Engineer working for a large technology conglomerate in HCM City, said that installing OpenClaw and connecting to the chatbot service is relatively easy, which can be done by non-tech users.
 |
|
Installing OpenClaw is not difficult, but it's more complex than conventional software. |
If not technically savvy, installing on a personal machine may entail some system or data-related risks.
If used locally, individuals follow the official documentation of OpenClaw and can use models from Ollama (there is a limited free version).
Meanwhile, Pham Minh from HCM City also said that installing OpenClaw is not difficult. Initial setup was simple, but in the process of loading third-party API services, connecting chatbots or adding skills was quite complicated.
According to this user, after connecting OpenClaw's trial with the Gemini 2. 5 Flash API, its "quota" quickly ran out after a few chats.
Luân Ngô argues that paying to install OpenClaw on a trial basis is unnecessary. However, as the task becomes more complex, owners must consult additional online tutorials and learn operating theory themselves.
OpenClaw is not completely free either. When running heavy-duty tasks such as image processing, data synthesis or multi-step automation cycles on a large scale, the cost of tokens through APIs increases rapidly.
To do more tasks, including the operation of LLM (large language model) investment at least 20-40 million VND, Luân said. This part of the cost does not take into account the installation of circuitry and transmission lines to ensure stable, continuous operating system.
The problem economic
OpenClaw is provided free of charge under the MIT License. However, this tool actually acts like an operating brain. To really deliver applications, you still need the Underlying Engine. The platform offers many options to go through servers of cloud processing services (API) such as ChatGPT, Claude or Gemini.
According to many users, OpenClaw generates far more tokens than traditional chatbots. As it works, the screenshots that the agent collects must be sent through a computer vision analysis model, causing the amount of processing resources to increase exponentially. At an enterprise scale, the total cost of the API combined with server infrastructure can range from $100,000 to $500,000. [citation needed]THREE HUNDRED AND FIFTY DOLLARS. every month for high-intensity operating systems.
 |
|
OpenClaw still needs strong hardware or large amounts of money paid to the AI server in order to operate. |
I used Anthropic's API and it cost me about 10 euros for about 15 minutes. When trying to run Ollama locally, the device was too slow to use. Switched to GLM's free API, but the machine overloaded, took about 7 minutes to get a response.
I'm not trying to build a sophisticated automation system or program agents. The workflow is very simple, one Reddit user complaining about the cost of setting up OpenClaw.
Below the post, some users commented that they were able to maintain stable performance with small language models such as Qwen3. 5 27B or Devstral-small-2:24B running locally.
See also: https://znews.vn/vo-mong-openclaw-post1640696.html
Comment (0)