Collection. Review. Production.
From the first collection to the final production.
Three tools. One clear path through the evidence.
All on computers you control.
Local processing. No outside AI service.
Follow the evidence
Scroll through the journey from scattered files to selected documents.
Skip the animation ↓Mail stores. Nested attachments. Scanned pages. Bring the scattered evidence into a searchable collection, with every unreadable file accounted for.
Search by keyword or meaning. Code, redact, and build a production with a privilege log. Keep a record of who made each decision and what it changed.
Ask questions in plain English, with links to the documents behind each answer. Explore who communicated with whom, with AI running on machines you control.
Move from a large collection to the documents selected for production. DFYCollect and DFYReview support production; use the tools independently or together.
Evidence, not promises
Discovery goes wrong when documents slip through: a scan nobody could search, an attachment buried three layers deep, a gap nobody noticed until the other side did.
Here is DFYCollect partway through the public Enron email collection, on one computer.
Look under the hood ↗The three tools
Use one, two or all three. DFY stands for Discover For Yourself.
Gathers and produces the evidence, including the documents other tools skip.
What it does
Review and produce on the computers your team already has.
What it does
AI-assisted review, running on machines you control rather than in the cloud.
What it does
Side by side
Each one works on its own, and each hands its work to the next.
| DFYCollect | DFYReview | DFYAI | |
|---|---|---|---|
| Status | Available now | Early access | In development |
| For | Collecting, searching and producing the evidence | Reviewing and producing it on your own computers | Reviewing it with AI to help |
| Highlights | Opens mail stores, zip files and documents inside documents. Reads scans. New search terms without starting over. | Loads any production. Searches by keyword or by meaning. Codes, redacts and produces, with a privilege log. | Plain-English questions, with links to the documents behind each answer. Who communicated with whom. |
| Runs on | Your computers | Your computers, shared over the office network | A review appliance we supply and maintain, or we stage the review for you |
| AI | None | No generative AI; search by meaning uses a small model on the computer's own processor | Models running on local machines; nothing is sent to an outside AI service |
| Explore DFYCollect → | Ask about early access → | Ask about early access → |
What you can stand behind
A search that quietly skips four thousand unreadable files looks exactly like one that read them all, unless the software tells you.
Every file that couldn't be read is listed with the reason, so you find the gaps before the other side does.
In DFYCollect, each document records why it was picked: its date, its location or its words.
If something would make the results incomplete, DFYCollect stops and says so rather than handing you a set that only looks finished.
All three run on local machines, never in a public cloud, and none sends a document to an outside AI service.
DFYReview and DFYAI record every coding decision: who made it, when, and what it changed. A batch coded in one step keeps the search that picked it.
An administrator can turn on the access log for a matter, and DFYReview and DFYAI then record who opened each document and when.
Every product page has a section called Current limits, in plain words. It's the part most vendors leave out, and the part you need most when deciding.
Changing the terms of the original collection means running it again. Searching the finished collection again does not.
Read DFYCollect's limits →No generative AI: nothing summarises, answers questions or writes for you. Search by meaning runs on the computer's own processor.
Read DFYReview's limits →Reading a production for people and facts is slow: about twenty-two days of GPU time for 31,000 documents, so it is run over a chosen scope.
Read DFYAI's limits →Try it
Bring a production, or let us collect one, and we'll show you what each tool finds, on computers you choose.