Chainlocker helps law firms run their document review playbooks in a standardized, repeatable way. Give it the matter's criteria, point it at the production, and every document comes back tagged, with written reasoning for what was found and a confidence score. It takes the routine first pass off a junior's desk. It does not give legal advice: every judgment stays with the lawyer.
The tags you want applied and the instructions for applying them: the clauses to find, the phrases to flag, the categories a document can land in. Saved per matter, reusable across matters.
A shared folder of PDFs, Word files, spreadsheets, email exports, and scans. It lists every file it understands and shows the count before anything runs.
Each document gets a tag from your list, written reasoning that says what was found, and a confidence score. Anything below the floor you set is routed to review, not filed.
Accept by default, reject with a correction. Working through a thousand documents this way is fast, and every verdict is recorded with who decided and when.
An Excel index with a clickable link to every file, plus an offline package that opens straight off a laptop in a courtroom with no wifi.
The criteria list is the standard. Run it again on the next production and the review is done the same way, whoever sits in the chair.
A playbook in Chainlocker is a list of tags and the instructions for applying them. Nothing about it is hardcoded, and nothing is supplied by us. The example here is illustrative, a shape rather than a recommendation. Your practice group writes the real one, once, and it becomes the standard every reviewer runs.
We'd like to see how your firm's playbooks are actually written. If you review contracts, productions, or diligence rooms and can spare an hour to walk us through what juniors are asked to look for, that shapes what we build.
Talk to us about your playbook →Chainlocker runs on a dedicated appliance installed at the firm and is reached from a browser on your LAN or VPN. Documents never leave it. AI inference runs on hardware we own over an encrypted private link, never a third-party AI API, so there is no usage meter and no provider terms of service between you and your client's files.
The deployment model in detail →Chainlocker is in development, targeting late 2026. Before it ships we want practicing lawyers running their own playbooks through it and telling us where it falls short. If that's you, or someone at your firm, we'd like to hear from you.