Hugging Face released relore on September 21, 2026, saying it helps AI agents track GitHub discussions and connect them to code changes. It is the company's first major update to its repository memory system since the introduction of Serge, its GitHub-native code review agent.
Hugging Face reported that relore indexes repository history and tracks who said what, treating maintainer decisions differently from contributor claims. The tool excludes machine-generated content from the default evidence set, ensuring more accurate context.
relore is built on GitHub repository data and targets improving the ability of AI agents to understand and review code changes. Availability begins with internal use across Hugging Face projects, including Transformers, and is expected to expand to external users in the future.
"Plain search turned out not to be enough," said Tarek Ziadé, Hugging Face contributor. "A maintainer saying 'yes, do this' is different evidence from a contributor guessing at a cause, or a bot leaving an automated review."
The announcement follows the development of Serge, Hugging Face's GitHub-native AI code review agent. Hugging Face said the goal of relore is to give an agent access to the same trail of decisions that maintainers carry in their heads while working on the current codebase.
Hugging Face did not say how relore will be made available to external users, and raised the open question of how to connect discussions to the code as it exists now. The company said it will continue to refine the tool and expand its use.
Source: huggingface
