Mage is an AI-powered legal diligence platform built for M&A and venture financing transactions. Key features and functions include:
Full Data Room Review
Mage reviews the entire document population in a data room, including the long tail of agreements that a sampled review would typically miss, rather than a selected subset of material contracts.
Document Classification and Linking
Documents are automatically classified and grouped into agreement families, with amendments, restatements, consents, and side letters resolved to their parent agreement.
Provision Extraction and Issue Flagging
Key provisions and data points are extracted into structured, comparable grids. Issues are flagged by severity and presented as an issues list or diligence memo, with a gap analysis across the data room.
Source Verification
Findings are cited to the specific page and clause of the governing document, allowing a reviewer to move from a flagged finding to the underlying language directly.
Disclosure Schedules and Variance Detection
The platform generates disclosure schedules from deal documents, reviews counterparty markups, and detects variance across agreements drawn from a common form.
Cap Table Tie-Out and Redlining
Mage ties out cap tables against equity documents and compares redlines across document versions.
Deal Deliverables
The platform produces request lists, management questionnaires, closing checklists, and drafted counsel, investment committee, and underwriting memoranda based on diligence findings.
Data Room Product
Mage offers a standalone data room for fundraising and M&A, allowing users to drag in files and folders, sync cloud storage (Box, Dropbox, Google Drive, OneDrive, Common Paper), or push documents via CLI, API, or AI agent. The room tracks a diligence readiness checklist, gates guest access behind NDAs and watermarking, and provides analytics on document engagement.
Developer and Agent Tooling
An open-source, MIT-licensed CLI (published to npm) and an MCP server allow the data room to be scripted from a terminal or operated by AI agents such as Claude or Cursor.
Loading...