The nonprofit Git hosting platform Codeberg for free and open-source projects, powered by the Forgejo software forge, now bans repositories that consist mainly of code written by generative AI.
The new restriction was added to Codeberg’s Terms of Use following a proposal discussed and approved through the organization’s community governance process. It applies to projects whose code is predominantly generated by AI tools, including services such as Claude and OpenAI Codex.
Codeberg states that such projects raise unresolved copyright issues, as the legal status of AI-generated code and its training data is unclear. The platform also notes that generative AI tools lack adequate safeguards against producing harmful or malicious code.
“You must not share projects that mostly consist of code written by “generative AI”-tools (including services such as Claude, OpenAI Codex). Such projects having an unclear copyright status (see requirements § 2 (1) 1 and § 2 (1) 3) and furthermore have little safeguards to ensure that they do not include harmful code (c.f. § 2 (1) 5).”
Importantly, the rule does not prohibit every use of AI in software development. Developers can still use AI tools for limited assistance, but repositories in which most of the code was generated by such systems are no longer permitted on the platform.
Notably, the wording does not provide an exception for projects whose AI-generated code has later been reviewed by a human. The deciding factor appears to be whether generative AI produced most of the code, rather than whether the resulting project was subsequently checked, understood, or maintained by its author.
At the same time, Codeberg has not explained precisely how it will determine whether a repository is “mostly” AI-generated, nor how the restriction will be enforced in practice. Reliably identifying AI-written code remains difficult, especially when generated output has been edited or mixed with human-written components.
In any case, this move places Codeberg in sharp contrast with GitHub, which is increasingly integrating AI coding assistants and autonomous agents throughout its platform.
This policy is among the clearest platform-wide restrictions against predominantly AI-generated software. It is not simply discouraging AI contributions, but requires that projects be hosted elsewhere.
