Multilingual Corpus and Moderation (MuLCAM)

MuLCAM is an evolving, community contributed corpus of slurs, derogatory terms and harassment patterns in Devanagari Nepali, Romanized Nepali and other mother tongues spoken in Nepal. It exists because TFGBV is widely under documented in local languages. Platforms and developers building moderation systems simply don't have local, multilingual data to detect it, and survivors rarely have structured evidence to point to.

The project works in three steps. Community members and partner organisations contribute text data, a feminist review process validates it for accuracy, and researchers, moderators and platforms use the resulting corpus for detection research and applications. Ethical detection and moderation tools, built as plug-ins, APIs and reporting layers, are in active development on top of the corpus. The work is grounded in the Feminist Principles of the Internet: access, rights, openness, and consent first, intersectional design.