Legal intelligence for complex cases.

Makor turns email, WhatsApp, call recordings, documents and scans into a verified chronology, an evidence matrix, and a factual record you can check. SOURCE · VERIFIED

Our only test is that you can check every word the system wrote and find it in the source. Anything that fails that is flagged, not deleted.

The problem

Complex cases are buried inside scattered material.

A serious commercial dispute arrives as thousands of emails, PDFs, contracts, spreadsheets, pleadings, invoices, scans and chat threads. The facts that decide it sit in disconnected sources, in different formats, in different mailboxes.

Associates and interns spend hundreds of hours building chronologies, evidence tables and document indexes by hand. The work is slow, expensive, and exposed to the one detail nobody caught.

0 documents in a typical evidence-heavy matter
0 of the early work is finding and organising, not lawyering
1 missed email is enough to lose a factual argument

Sources

The case does not live in email alone, and it is not waiting tidily in a folder.

A real case lives in WhatsApp threads, call recordings, scans of old documents and spreadsheets with hidden rows. Makor ingests all of them, spots duplicates and versions, and links every item to the facts it proves. It also reads what was not meant to be read: hidden spreadsheet rows, deleted comments and version history.

  • GmailWhole mailboxes, attachments and threads included
  • Google DriveShared folders, versions and permissions
  • WhatsAppFull chat exports: text, media, groups
  • Call recordingsTranscribed and cross-checked against documents and email
  • PDFIncluding scanned Hebrew PDFs, quoted exactly
  • WordContracts, letters, drafts and version comparison
  • ExcelIncluding hidden rows and hidden comments
  • ScansHebrew OCR for old documents and faxes
  • Court filingsPleadings, rulings and transcripts
  • Mail archivesEML / MSG / whole ZIP archives

What someone tried to bury becomes a finding.

Product

The factual record your team builds by hand, already built.

Point it at the raw material of the case. Makor ingests everything, extracts the facts, links each one to its source, and assembles the structured litigation file the team would have built by hand.

It replaces the first heavy layer of manual factual review, so the lawyers start from an ordered, verified file instead of a folder of chaos.

  1. 01
    Connect the sourcesMailboxes, drives, WhatsApp, recordings, folders.
  2. 02
    Ingest the documentsEvery format, every attachment, with duplicate detection.
  3. 03
    Extract entities and datesPeople, companies, amounts, deadlines.
  4. 04
    Build the factual timelineOne chronology from every source.
  5. 05
    Link every fact to its evidenceExact quote, page and clause.
  6. 06
    Surface contradictions and gapsInconsistencies and missing documents flagged.
  7. 07
    Map the cast of charactersWho said what, to whom, when. Built from the correspondence itself.
  8. 08
    Coverage reportEverything no case theme touched, listed.

Inside the workspace

A litigation file you can interrogate.

MATTER WORKSPACE · DEMO DATA
12 January 2023
Framework agreement signed between Almog Development and Meridian Construction. Delivery set for 30 June 2024. EX-7 · CONTRACT · §4.2
3 August 2023
First notice of delay. The site manager reports a nine-week slip on the frame over WhatsApp, blaming a supplier failure. EX-142 · WHATSAPP · 14:32
15 September 2023
Progress report claims the schedule is on track. It contradicts the same team's internal report of 3 August. CONTRADICTION · EX-198 vs EX-142
2 February 2024
Payment withheld. Milestone 4 invoice disputed; ILS 2.4m held pending approval. EX-311 · INVOICE + EX-314 · LETTER
FactSourceWhoDateConfidence
Delivery date set for 30 June 2024EX-7 §4.2Both parties12.1.23Proven
Nine-week slip reported internallyEX-142Site manager3.8.23Well supported
Supplier failure caused the delayEX-142, EX-155Contractor8/23Partial
The developer knew of the delay before the September reportInferenceInference
The schedule was on track in September 2023EX-198Contractor15.9.23Contradicted

Who knew about the delay before the September progress report?

The site manager reported a nine-week slip on 3 August 2023, six weeks before the progress report. It was sent in a WhatsApp group that included the developer's project manager.

F-0142 · VERBATIM VERIFIED EX-142 · WHATSAPP · 3.8.23 14:32 EX-155 · INTERNAL EMAIL

Did the developer approve an extension of the delivery date in writing?

I have no sufficient evidence in this file. No document was found in which the developer approves an extension. Searched: extension of time, schedule approval, delivery deferral, additional days. Zero verified hits.

GAP · MISSING: WRITTEN EXTENSION APPROVAL

Makor answers only from what is in the file. When there is no evidence it says so and shows what it searched. Every quote in an answer is checked against its source before it is shown.

ParaPlaintiff's allegationForAgainstStatus
§12Delivery was eleven months late with no prior noticeF-0142, F-0155F-0201Mixed
§14The defendant never reported any schedule slippageF-0142Contradicted
§19The plaintiff paid the full consideration for milestone 4No evidence
§23Written notice was served on 2 February 2024F-0311Supported

The complaint goes in and every paragraph is mapped to the evidence on file. Each paragraph's quote is checked against the pleading itself. The paragraphs with no evidence are what to chase in discovery.

How it works

What happens between the pile and the research file.

Input

  • Gmail
  • Google Drive
  • WhatsApp
  • Call recordings
  • PDF
  • Word
  • Excel
  • Scans
  • Court filings

Processing

  • Hebrew OCR
  • Call transcription
  • Entity extraction
  • Date extraction
  • Email thread analysis
  • Document classification
  • Duplicate detection
  • Fact extraction
  • Source linking
  • Contradiction detection
  • Confidence grading

Output

  • Factual timeline
  • Evidence matrix
  • Disclosure list
  • Pleading analysis
  • Cast of characters
  • Missing documents list
  • Analysis coverage report
  • Verified-quote rate

Capabilities

Nine instruments. One verified file.

Quote verification

Every quote is checked character by character against the document's own text layer. Anything not found as written is flagged, and the fact resting on it is downgraded.

Factual timeline

The full chronological story of the case, with every event linked to its evidence.

Evidence matrix

Scattered documents become a structured table of facts, sources, people and dates.

Ask the case

A question in plain language, an answer built only from the evidence, every source one click away. When the file holds no answer, it says so instead of guessing.

Contradiction detection

Inconsistencies between emails, contracts, statements, timelines and filings, surfaced automatically with both quotes side by side.

Pleading analysis

The complaint goes in and every numbered paragraph comes out mapped: what proves it, what contradicts it, and what has no evidence at all.

Cast of characters

The people list is built from the correspondence itself: addresses, nicknames, WhatsApp handles and spelling variants. Duplicate identities are proposed for merging.

Disclosure list

A draft disclosure inventory from everything ingested: numbered documents, email threads, WhatsApp, privilege-suspect items flagged for review, and an annex of evidence gaps for a specific-disclosure request.

Analysis coverage report

The list of documents no case theme ever touched. The analysis names its own blind spot instead of hiding it.

The advantage

The system reads Hebrew, scans included.

Foreign systems translate. Makor thinks in Hebrew: real Hebrew search that understands prefixes, inflections and spelling variants; exact quotes out of scanned Hebrew PDFs; and right-to-left deliverables in legal Hebrew, ready for an Israeli court.

  • Hybrid Hebrew search across the whole file, transcripts included
  • Exact quotes with page and clause, scans included
  • Typeset RTL deliverables, ready for filing and hearing
  • The material stays in a private, isolated environment

Trust and verification

Every quote is checked against its source before anyone sees it.

Every quote that enters the file is checked automatically against the source itself, character by character, against the document's own text layer. A quote that is not found as written is not deleted quietly: it is flagged and the fact resting on it is downgraded. At the end of every analysis there is one number you can show a partner, the percentage of quotes that verified.

Verbatim verified

The quote was found in the source exactly as presented. This is the grade you can take into a hearing.

Paraphrased, labelled

The meaning survived but the wording is not identical. The item is labelled and never shown as a verbatim quote.

Not found in source

The quote does not exist in the document. The fact is downgraded automatically and reaches the lawyer flagged.

Proven by document

A verbatim quote, in the source, at the cited location.

Well supported

Several consistent sources point the same way.

Supported but incomplete

Evidence exists; a link in the chain is missing.

Inference only

Reasoned from the record and marked as such. Never silently.

Contradicted / problematic

Sources disagree. Flagged for the lawyer's attention.

And what was never examined: Makor produces the list of documents no case theme touched. The analysis names its own blind spot instead of hiding it. Makor stops at the facts layer. It does not draft claims, write documents or propose strategy. That is your work.

Security and deployment

The file stays on your servers.

Makor is installed at your firm: on the firm's own server, inside your network. There is no cloud of ours for client material to pass through, and nothing reports back to us. The licence is digitally signed and checked locally, with no internet connection required. The system carries the firm's name.

  • Installed on the firm's serverYour environment, your backups. Client material never leaves.
  • Standalone licenceChecked locally, with no call home and no remote monitoring.
  • Firm-owned AI keyUsage is metered per matter and stops at the budget you set.
  • Matter wallNeed-to-know access only. Even an administrator gets no silent access to content.
  • Two-factor authenticationTOTP, lockout after failed attempts, and sessions that expire on their own.
  • Full audit logEvery view, every change, every break-glass. Stored and read at your firm.
  • Attorney gateNo analysis runs until a lawyer has approved the case setup.
  • White labelThe firm's name and colours. To the client it is the firm's system.

Where this fits

Built for evidence-heavy matters.

  • Commercial litigation
  • Real estate disputes
  • Construction delay claims
  • Contractor disputes
  • Shareholder and indemnity disputes
  • Insurance subrogation
  • Evidence-heavy civil litigation
  • Internal investigations
  • Pre-action case assessment

The change

What actually changes.

Before Makor

  • Thousands of files across mailboxes and drives
  • Email and WhatsApp threads scattered and unsearchable
  • Chronologies built by hand in a spreadsheet
  • The risk of a missed exhibit in every matter
  • Slow, expensive document review
  • Sources that are hard to trace back

After Makor

  • One structured matter workspace
  • A clickable evidence link on every fact
  • A verified factual chronology
  • The cast of characters in the matter
  • A missing-documents list, produced automatically
  • Lawyers start on strategy, not on sorting

Book a demo

See how Makor turns case material into litigation intelligence.

A 30-minute demo on an anonymised matter. Bring the hard kind of case you deal with, and we will show you what it looks like from the other side.

No client material is ever used in demos.