Lemmatic Authorities

Paste a raw export into a chat and the model reads what fits, then bluffs the rest. Authorities gives it cases it can read to the end.

One Markdown file per case, plus a manifest your model reads first — the court's text, complete, with the page numbers you cite and token counts up front.

Files convert in your browser and are never uploaded. The page can't send your files, or anything about them, anywhere: its security policy, enforced by your browser, lets it fetch its own files from our server and nothing more.

Files

Drop RTF files here, or choose files. Westlaw Find & Print RTF only — anything else is declined, not half-converted. Several files at once, or one .zip of them.

Loading…

Options

What to keep The Attorneys and Law Firms block Official-reporter headnotes, in a labelled section Mark the opinion modifications in the text

How to divide it

Numbered volumes instead of one file per case Tokens per volume

Split a case that is over the threshold into parts Oversize threshold

Report

Release checks

Download the zip. If your AI tool opens archives (ChatGPT does), attach the zip as it is; in a ChatGPT Project, add it to the Project’s files and replace it there when you make a new one. If your tool doesn’t open archives (Claude doesn’t), unzip it and attach manifest.md first, then the case files, or the volumes if there are any.
Paste the instructions as your first message, or as the Project’s instructions. They tell the model to read the manifest first and to confirm each case ID before summarizing it.
Before anything the model gives you goes into a brief, cite-check it against the original authority. The manifest’s “Check these” list is where to start, not where to stop.

Converting more exports? Drop them all at once and they become one bundle, numbered straight through. Converted separately they are two bundles, and their files have the same names.

Download

I have a key

Bundle

File	Bytes	Copy
Total

Reference

Getting the export

In Find & Print, choose RTF as the format and deliver all the cases as one file. More than one export? Drop them together — they become one bundle.

What's in the files

Everything in the files is the court's own text as published — captions, opinions, footnotes, and star pagination. Publisher summaries and annotations aren't part of the opinion and aren't included. Copyright and source notices from the export are preserved in each file's front matter.