DeepSeek has released an open-source prototype of a more autonomous coding agent, called Super-agent, Bloomberg reported on Sunday. According to Bloomberg, the release is aimed at competing with Anthropic’s Claude Code.
The Chinese AI company’s Harness team described the agent in a WeChat post on October 31 and published the code on GitHub .
The agent can watch developers’ group chats and spot possible bugs even when no one tags it, according to the post. It then tries to reproduce the problem and work on a fix in the background, Chinese outlet Yicai reported .
The team said the aim is to take up as little human attention as possible. DeepSeek already runs Super-agent internally to fix bugs, respond to code reviews, and improve its own tools, according to the post.
However, the project’s GitHub page describes it as a research prototype for reference only, not an official DeepSeek product. It also advises developers to review and adapt the code before using it.
The prototype is released under the MIT licence, which lets anyone use and change the code. It works with GitHub, with optional support for Feishu, the workplace messaging app owned by ByteDance.
According to the project’s documentation, the agent only starts work when an approved person asks it to, or when a rule set by the team allows it.
Super-agent runs on DeepSeek Harness , the company’s open-source software that wraps its AI models so that agents can handle longer and more complex tasks. DeepSeek formed the Harness team earlier this year to build agents that can compete with coding tools from US labs, Bloomberg reported.
The release follows a $15bn funding round that Bloomberg reported in October, with Tencent and battery maker CATL among the investors, ahead of a possible stock market listing. DeepSeek has not confirmed the round.
Get the most important tech news in your inbox each week.
