The Good Janets is a project of the SIGNAL Network and is funded by the Social Sciences and Humanities Research Council.
Additional financial and in-kind support comes from:
The Good Janets gathers transcripts from a growing list of progressive podcasts. These transcripts provide useful counterpoints to those included in the ManoWhisper database and include social justice oriented content that explores equity; centers and amplifies underrepresented and targeted people including women, 2SLGBTQ+, Black, Indigenous, and racialized communities; or tackles pressing political, cultural, and/or technological debates. This endeavour aligns with SIGNALâs broader research goal of mapping the intersections of digital, political, and cultural gender discourse and the ecosystems that support resistance against the increasingly proscriptive discourse of gender conservatism, male supremacy, white supremacy, and facism.
This site is built with Flask, with search powered by Apache Solr, and has an API.
| Method | Model / Tool | Notes |
|---|---|---|
| Transcription | Whisper (turbo) | |
| Misogyny classification | MilaNLProc/bert-base-uncased-ear-misogyny | Indicator, not a definitive judgment. |
| Toxicity classification | s-nlp/roberta_toxicity_classifier | Indicator, not a definitive judgment. |
| Hate speech classification | facebook/roberta-hate-speech-dynabench-r4-target | Indicator, not a definitive judgment. |
| Episode summaries | gmurro/bart-large-finetuned-filtered-spotify-podcast-summ | Can miss key points, hallucinate, introduce inaccuracies, and/or unintended comedy. |
| Host/guest/mentioned name extraction | spaCy (en_core_web_sm), reconciled against Wikidata | Automatic extraction; expect occasional errors and false positives. |
| Topic extraction | Qwen2.5-3B-Instruct | Automatic extraction; expect occasional errors and false positives. |
Show and episode statistics, along with the classifications above, are generated with a companion command-line tool called manoclassify. This website grew out of a set of scripts developed during a 2024 sabbatical project.
Users should approach all of the above critically, treating them as indicators rather than definitive judgments. In our own work, we verify computational findings through close reading of transcripts, situate patterns within broader political and cultural contexts, and remain attentive to the limitations of any automated classification system.
How to cite the ManoWhisper or The Good Janets projects in your research:
Ruest, N. et al. (2026) "Infrastructures of Listening: The ManoWhisper Podcast Analysis Pipeline," Digital Humanities Quarterly, 20(1). Available at: https://doi.org/10.63744/egsbwnv6t5c6.
Your citations help to further the recognition of using open-source tools for inquiry, assists in growing the research community, and acknowledges the efforts of contributors to this project.