Digital files, scans, handwriting, any language, last year or last century. ArchiveLM finds what you're looking for by meaning, not just keywords. When it can't read something, it says so. It never makes something up.
The answer is usually somewhere in the file: an email saved as PDF, a clause in a contract, a memo from 1962, a handwritten ledger, a Spanish-language version. Keyword search misses different wording, misspellings and other languages, so teams still hunt folder by folder. Miss the one page that shows who knew what, and when, and the case turns on it.
Who knew about the defect, and when? Show me every document, oldest first.
Find the clause where the supplier accepts liability, in the English or the Spanish versions.
Which people and companies appear around the 2014 sale of the Ridgeway site?
Upload the matter's files. Digital documents are read directly; scans and handwriting are read by AI, with handwriting read twice and cross-checked.
Every page is indexed by meaning, along with the people, companies, places and dates it mentions, in any language.
Search or ask in plain English. Answers come from your documents only, with the source page cited.
One click opens the exact page behind any answer. Pages that couldn't be read reliably are flagged, never guessed.
Find it in minutes, not days: one search across the whole matter, digital and paper, old and new
Answers you can stand behind: every answer is tied to a page you can open, so nothing invented reaches a brief
Finds what keyword search misses: different wording, misspellings in old scans, another language
Timelines, who-knew-what maps and cited findings drop straight into memos, deposition prep and expert reports
In the works: research beyond your own files (historic US newspapers via the Library of Congress, cited web research and, planned, published case law), always labelled and kept separate from the matter's documents
Each account's documents sit in an isolated database and storage bucket, enforced by row-level security — no other customer can see or search them.
Encrypted in transit (TLS 1.2+) and encrypted at rest.
Your documents are never used to train AI models — we instruct our providers not to train on your data.
Documents are stored in the United States.
Deleted on request, and removed within 30 days of an account closing.
A published list of every subprocessor that touches your data, with its role and location (legal.archivelm.com/legal/subprocessors on the main site).
We are not yet SOC 2 certified. We'll share our security overview and subprocessor list on request.
| Plan | What's included | Price |
|---|---|---|
| One-off | ||
| Pilot | A sample of your own documents (a closed or non-privileged matter is ideal), searched and walked through with you | Free |
| Matter setup | We load and index the matter: up to 25,000 digital pages, or 5,000 scanned pages, or 1,500 handwritten or foreign-language pages. Checked with your team, plus a 30-minute walkthrough | $2,500 per matter |
| Additional pages | Digital / scanned / handwritten or foreign-language | $0.01 / $0.40 / $1.50 per page |
| Ongoing | ||
| Active matter | Up to 10 users on the matter team, all search and research tools, exports and support. Pause or close any time | $1,250/month |
| Additional users | Beyond 10 on a matter | $99/user/month |
| Firm plan | Up to five active matters and 50 users. Matter setups and extra pages billed as above | $4,500/month |
| Security | ||
| Standard security | Encryption, account isolation, no AI training on your documents, deletion on request | Included |
| Private environment | A dedicated database and storage on your own web address, shared with no other customer. (Isolation, not a compliance certification.) | $7,500 setup + $1,500/month |
For perspective: a typical six-month matter (setup plus six months) is about $10,000, a fraction of the associate and paralegal hours spent hunting through a large document set by hand.
During our beta, matters are sized up to 100,000 pages; larger collections on request.
Having documents in digital form isn't the same as being able to find things in them. Keyword search misses different wording, misspellings and other languages, and it can't answer a question. ArchiveLM searches by meaning across every document in the matter and answers with the exact page it relied on. Digital files are read directly; scans and handwriting are read too, so paper and digital sit in one search.
No. There's no legal hold, Bates numbering or privilege review here. ArchiveLM is the search and research layer you use alongside your review platform, to find and understand what's in the documents.
Not yet. That's in the works: searching public sources such as historic US newspapers through the Library of Congress, cited web research and, planned, published case law. Outside results will always be labelled and kept separate from your matter's documents.
It's flagged rather than guessed. A page that can't be read reliably is marked for review instead of producing a confident-sounding answer that isn't actually on the page. That's the difference between a citation you can rely on and one you can't.
No. We instruct our AI providers not to train on your data, and each client's documents are stored in an isolated account that no other customer can see or search.
Yes. A private environment gives you a dedicated database and storage on your own web address, shared with no other customer. It's $7,500 setup plus $1,500 a month, on top of standard security. It's isolation, not a compliance certification.
Give us a sample from a closed or non-privileged matter, digital files plus your hardest scans, and we'll show you what ArchiveLM finds, how it cites it, and where it can't read a page. No cost, no commitment.
ArchiveLM is in private beta. We review each request and typically respond within 1–3 business days.