Moderation Process

Every contribution goes through a two-stage automated pipeline before it can become published study content. The first stage is content extraction, where the system transforms your uploaded file into structured text. For documents, this means pulling out the text page by page, identifying tables, and extracting embedded images. For images and scanned documents, an AI vision model classifies and reads the visual content.

Once extraction is complete, the structured text enters the second stage — the AI Teachers' Council. This is a five-model review pipeline where each AI teacher has a specific role. The first teacher, Luna, acts as the Creator and generates initial study content from your source material in real time. The second teacher, Atlas, serves as the Reviewer and evaluates the content for accuracy, completeness, and pedagogical quality. The third teacher, Nova, is the Enricher, adding examples, deepening explanations, and refining the language.

The fourth teacher, Sage, acts as the Validator, checking factual accuracy, internal consistency, and fidelity to the source material. The fifth teacher, Echo, is the Fact Checker, cross-referencing claims against real-world knowledge for a final verification pass. Each teacher reviews the content independently without seeing what the others have said, creating a genuine peer review process. When all five teachers agree, the content is verified and published.

If any teacher identifies issues, the content is flagged for further review. Contributions that are rejected include a reason explaining what went wrong, and you can submit an appeal if you believe the rejection was incorrect. The entire process is designed to ensure that only accurate, high-quality study material reaches other students.