Claim Ledger Lite is a small local academic-writing helper. Paste a draft paragraph or study note, paste evidence records, and it turns the text into a claim-by-claim review ledger.

It is built for one narrow job: make unsupported claims visible before they become polished prose.

Live demo: https://bte808.github.io/fun-20260523-c-claim-ledger/

• Split a draft into reviewable claims.
• Parse simple evidence records such as E1 | label | excerpt.
• Detect explicit evidence links like [E1].
• Suggest candidate evidence when a claim and evidence note share key wording.
• Flag strong claims that have no visible support.
• Surface review warnings for duplicate IDs, thin excerpts, placeholder evidence, and over-strong claims.
• Filter the review ledger by linked, candidate, needs-source, or context rows.
• Export a Markdown table for a paper checklist, lab notebook, study guide, or reviewer comment.
• Copy a focused review plan that lists only claims and evidence records needing source cleanup.

• Checking a literature-review paragraph before adding it to a paper draft.
• Turning lecture notes into a source-aware revision checklist.
• Reviewing lab-report conclusions against observation notes.
• Marking which claims in a study guide are source-backed and which are interpretation.

Academic notes often fail quietly: the prose sounds reasonable, but the evidence path is unclear. This tool does not judge truth. It forces a concrete workflow:

1. Name the claim.
2. Point to evidence.
3. Mark candidate matches.
4. Rewrite or soften unsupported statements.

That makes paper reading, course revision, and lab writing easier to audit.

It treats citation hygiene as a visible ledger instead of a last-minute formatting task. The fun part is watching a paragraph become a dashboard of linked, candidate, and risky claims without sending text to any service.

• It is a small, local-first writing utility with no account, API key, or upload step.
• It turns vague "check the sources" work into a concrete action list.
• It is useful for papers, lab reports, course notes, and study guides without pretending to verify truth.
• The code stays intentionally lightweight: static HTML, CSS, JavaScript, MIT license, and smoke tests.

This project was inspired by recent public discussion around human-in-the-loop academic writing workflows and evidence ledgers:

• Academic Research Skills for Claude Code, a recent public project around research/write/review/revise workflows and integrity checks: https://github.com/Imbad0202/academic-research-skills
• Open Knowledge Maps as an example of making literature work visible through maps: https://openknowledgemaps.org/
• Claim-Evidence-Reasoning teaching materials that frame academic explanation as claims supported by sufficient evidence and reasoning: https://aroles.github.io/biol211/articles/CER_framework.html

The implementation here is original and does not copy source code, design, protected text, papers, or data.

• It does not verify whether a claim is true.
• It does not fetch or summarize papers.
• It does not create citations.
• It does not replace textbooks, source papers, instructors, reviewers, or domain experts.
• The bundled sample is synthetic example data and must not be treated as a real citation or research conclusion.

You can also use the public demo:

https://bte808.github.io/fun-20260523-c-claim-ledger/

npm install
npm start

Then open:

http://localhost:5176/

There are no runtime dependencies. npm install only creates a normal npm project state; the app itself is static HTML, CSS, and JavaScript.

1. Paste draft claims or load the sample.
2. Paste evidence notes, one per line.
3. Use bracket IDs such as [E1] in the draft when a claim has support.
4. Read the Claim Review table.
5. Review any warning badges before reusing a claim.
6. Copy or download the Markdown export.
7. Use Copy Review Plan when you only want the next source-cleanup actions.

Evidence format:

E1 | Article note | The source says spaced practice supports later recall.
E2 | Lab observation | Trial 2 had lower error after calibration.

npm run check

The check runs syntax validation for the browser modules and a Node smoke test for claim parsing, evidence linking, status counts, and Markdown export.

MIT

• Import/export JSON sessions.
• Add a source-strength field such as primary study, review, lecture, or observation.
• Add per-claim confidence notes.
• Add a print stylesheet for paper-margin review.