DeepSeek's Harness Hits 1.54 Million Daily Users, Fueling the AI Agent Race

Growth sprint, measured in weeks

On Sunday, DeepSeek said on X that its open source agent app, Harness, reached 1.54 million daily active users as of November 4. That is less than three months from its August 13 launch.

The company added that by the end of September, daily users stood at about 271,000, coming immediately after a desktop version of the app was rolled out. These figures are derived from how many users connect to DeepSeek's own servers.

What Harness does and why it is heavy on compute

DeepSeek is targeting infrastructure for AI agents - software that performs tasks for a user with minimal oversight. These agents take on multi-step tasks such as coding and editing code, parsing files, running research, and operating additional tools.

That workload burns far more processing than a standard chatbot. DeepSeek says Harness is chewing through approximately 92 trillion tokens a day, roughly 60 million per active user, as agents run for hours reading and writing code, traversing tools, and browsing with limited human input.

New AI tools are reshaping how investors discover opportunities and manage risk. We're hosting our first ever investor livestream on 11/17 to show you how to use ChatGPT and Claude to invest like a pro, hosted by Briefs Finance CEO Jaspreet Singh. Save your seat.

Plugins everywhere and a marketplace on deck

DeepSeek says it is building Harness as an "everything is a plugin" architecture. A plugin adds a feature without changing the underlying app, extending what the software can do. The company counts about 35,000 plugins so far, and says 60% of users have created one or more, though sophistication varies.

To grow that ecosystem, DeepSeek intends to launch a formal marketplace, allowing developers to publish plugins and letting users discover them. The company also plans a support initiative for plugin creators, offering token grants, technical help, and promotion opportunities. The company hasn't yet shared additional specifics.

The competitive frame and the investor angle

The Chinese AI disruptor, which jolted Silicon Valley last year with its low cost reasoning model R1, is positioning Harness as an open, developer driven agent ecosystem aimed at challenging OpenAI and Anthropic PBC not only on cost and performance, but also on the range of applications created around its technology. The company is betting that third party applications and specialized tools built on its open source agent infrastructure can become a durable edge as the industry shifts past chatbots and toward autonomous agents.

Why does that matter? The battleground is shifting from model scores to ecosystems and integration, and that shift can influence which platforms attract developers, users, and ultimately capital - all inputs your portfolio should watch.

Understanding practical applications of emerging technology can give investors a meaningful advantage. On 11/17, our CEO Jaspreet Singh goes live for our first ever investor livestream on how to use ChatGPT and Claude to invest like a pro. Register to join him.