Verified English content. AI-assisted learning. Human-controlled quality.

Where verification has been applied

Sources we cross-check against

Where a reliable source exists, we use it. Where it does not, we say so. We never invent a citation to make analysis look better-supported than it is.

• Project Gutenberg - Public-domain editions of pre-1928 texts including Shakespeare, the major Victorian novelists, Romantic and Victorian poetry, and the prescribed nineteenth-century novels on most exam-board specifications. Quotations are verified against the Gutenberg text rather than memory.
• Folger Shakespeare Library - For Shakespeare texts we cross-reference the Folger digital editions because of their careful editorial apparatus. Where act/scene/line references differ between editions, we explain the discrepancy rather than pretend it does not exist.
• BBC and broadsheet archives - For context - biographical, historical, social - we cross-check against named published sources. We prefer the BBC, the major broadsheets and academic publishers over open-web summaries.
• Public exam-board specifications - Assessment objective wording, paper structure and prescribed-text lists are checked against the public specification PDFs published by AQA, Pearson Edexcel, OCR, WJEC Eduqas and Cambridge International. Where a specification updates, we update.

Confidence levels

Not all content carries the same weight of evidence. We use four internal confidence labels so a reader can tell what stands behind a specific page.

• Highest - quote checked + human reviewed

  Quotations cross-referenced against a public-domain source. Analysis read end-to-end by a human editor. Used for set-text pages and exemplar essays.

• High - human reviewed

  AI-drafted then read end-to-end by a human editor. Used for theme-level analysis, contextual notes and revision summaries.

• Medium - AI-drafted, awaiting review

  AI-generated content surfaced as a learning aid but not yet signed off by a human editor. We mark this clearly so students know what they are looking at.

• Formative - practice feedback only

  AI-generated essay feedback and grade indications. Useful for practice and pattern-spotting, but not official marking. Always to be discussed with a teacher before being acted on.

What we have caught and fixed

AI-assisted content gets two specific things wrong more than anything else: invented quotations that sound right but do not appear in the text, and confident statements about historical context that no source actually supports. We keep an internal log of corrections and bump the "Last updated" date on the affected page when a substantive change lands.

Spotted something that does not look right?

Report it and our team will review the page. Reports go to our editorial inbox, are logged within the working day, and the reporter is told the outcome once the review is complete.

Report a content issue →