9h ago

Evidence Ledger & Claim-to-Source Audit for Non-Fiction

I would love to see an Evidence Ledger or Claim-to-Source Audit for non-fiction projects. Scribeist already has strong research and citation features, but it would be extremely useful to have a higher-level system that connects substantive claims in the manuscript to the evidence supporting them. For example, Scribeist could identify factual or research-based claims and show something like: ✅ Supported — linked to one or more sources 🟡 Citation needed ⚠️ Evidence is weak or indirect 🔀 Multiple or conflicting sources 💭 Author interpretation / synthesis 📝 Personal experience or observation For each claim, it would be useful to see: - the exact manuscript passage; - linked supporting source(s); - the relevant quote, passage, page, section, or locator from that source; - citation status; - optional notes about the strength or relevance of the evidence; - conflicting sources where applicable; - whether the statement is a factual claim, interpretation, hypothesis, personal experience, or author synthesis. An “Audit Manuscript” function could then scan a chapter or entire book and provide an overview such as: - claims with citations; - potentially unsupported factual claims; - citations that do not appear to support the surrounding claim strongly enough; - research notes that are not yet used; - places where several sources disagree; - chapters with unusually low or high evidence coverage. The author should always remain in control. I would not want the AI to automatically decide whether something is “true,” but rather help surface the relationship between claims and available sources for human review. Being able to export the Evidence Ledger would also be extremely valuable for editors, fact-checkers, publishers, academic reviewers, and authors working on evidence-heavy non-fiction. This could turn Scribeist from a writing tool with citations into a genuinely powerful evidence-aware non-fiction writing environment.

Planned