FRCP Rule 26(a)(2)(B) · for engineering expert witnesses
The Rule 26 report tool that checks every citation before opposing counsel does.
A drafting tool for federal engineering expert witnesses. It structures Rule 26(a)(2)(B) reports from your working notes, validates every case and paper citation against authoritative sources, and keeps your cumulative CV and rolling four-year prior-testimony list current across every case.
Why this exists
Generic AI is now an active liability in expert reports.
Two recent federal rulings, both from 2025, struck expert filings for AI-hallucinated citations:
- Kohls v. Ellison, D. Minn., January 2025. A Stanford professor’s expert declaration cited two nonexistent academic articles and misattributed a third. The court struck the declaration entirely and barred the state from submitting a revised one.
- Concord Music Group v. Anthropic, N.D. Cal., May 2025. An expert declaration containing a hallucinated citation was struck. The court noted the AI-assisted manual citation check had failed to catch a complete change in article title.
For the expert, the cost is not the citation. It is the credibility — once the fact-finder is told the expert couldn’t get their own references right, the opinion is tainted, and a small market has a long memory.
rule26 is the disciplined alternative. Every case citation is checked against CourtListener; every paper citation against Crossref, Semantic Scholar, and OpenAlex. Anything unverified is flagged for your confirmation — never silently included.
How it works
Three things rule26 does that generic tools cannot.
Your cumulative state, always correct.
Your CV, your publications, and your rolling four-year prior-testimony list are maintained across every case and inserted into every report — automatically, with the four-year window recomputed against each report’s date. The list you enter once stops going stale between cases: the deposition-trap of citing an out-of-date window is taken off the table.
Every citation, verified before you sign.
Case citations are checked against CourtListener and confirmed by name-match — not just existence. Paper citations are checked against Crossref, Semantic Scholar, and OpenAlex. Unverified or mismatched references are flagged; you confirm each one before the report can be finalized.
You author. The tool drafts, structures, and checks.
Opinions come from your notes — the tool does not invent, strengthen, or add reasoning you did not give. Every finalized report is signed by you and carries the assistant-not-author disclosure required for the unsettled 2025-26 AI-authorship question.
Before you trust it with a live engagement
What drafts your report, and what happens to your case material.
What actually drafts your report.
A leading frontier model (Anthropic’s Claude) takes your own working notes and reformats them into the proper FRCP Rule 26 document structure — it never originates opinions or reasoning. It also reviews the draft and suggests places that may invite challenge; those are suggestions only, and it never alters your opinions or your reasoning. Citation validation is not the AI: every reference is checked by a direct lookup against the citation libraries. Your qualifications, CV, prior-testimony, and compensation sections are generated deterministically from the data you enter, with no model involvement at all. You remain the author of record.
Your case material, handled like case material.
Your notes are processed only by the drafting model, and they are not used to train any AI model. Citations are validated against authoritative public databases such as CourtListener, Crossref, Semantic Scholar, and OpenAlex — we send only the citation itself, never your case notes. Your data is isolated to your account by row-level security, and it is never sent to analytics or ad platforms.
The only thing we keep for the long term is a cryptographic fingerprint of your finalized report, together with its citation verdicts and your acknowledgements — not your case content. It is a Daubert asset, not a retention burden: it can show that a human authored the report and that the document you filed matches what you finalized, all without us holding on to your notes or your drafted report. Your notes and drafts live only while your account is active.
Pricing
$79 per month. One price.
Month-to-month, cancel any time from the customer portal. No contract, no lock-in.
One struck declaration can taint an opinion and follow you in a small market. At $79 a month against a $300–600/hr practice, it’s the cheapest line item you carry.
- Federal FRCP Rule 26(a)(2)(B) report drafting from your working notes.
- Cumulative CV, publications, and rolling four-year prior-testimony list, auto-maintained.
- Case-citation and paper-citation validation; unverified references flagged for your confirmation.
- Word (.docx) export of the finished, expert-signed report.
- An audit log of every change to your cumulative state.