Vquity 1.18 is out, and nearly all of it lands in one place: the Data Room. This release collapses four competing upload buttons into a single Add Documents flow, introduces a real document taxonomy of roughly 120 types across 8 sections, chains renewals into versioned series automatically, and (the part we care most about) makes every AI-extracted value editable in place, with a full audit trail behind each correction. Here's what shipped in the Vquity 1.18 release, why we built it this way, and where the Data Room goes next.
One way in: the Add Documents flow
Before 1.18, the Data Room header had four separate buttons (Extract with AI, Upload Doc, Import Folder, and a plain Upload) that did overlapping things through different dialogs. Four entry points meant four slightly different outcomes, and picking the wrong one meant redoing work. They're now a single Add Documents split button. The dropdown still reaches every path (file upload, whole-folder import, direct AI extraction, and manual record entry), but everything feeds one pipeline.
That pipeline classifies before anything uploads. Drop a set of files and each one is analyzed instantly: document type, status, executed and expiry dates, duplicate detection, and stakeholder suggestions, all presented in an editable review list. If the classifier calls your consulting agreement an employment agreement, you fix the type in the review list, before the record exists, not after it's been filed into the wrong section.
And once an upload completes, one click hands the same files to the universal AI extractor, which verifies their types and files key data (parties, amounts, dates) into the cap table. Upload and extraction used to be separate decisions made in separate dialogs. Now extraction is just the next button.
A real taxonomy: ~120 document types in 8 sections
Most document vaults ship a handful of generic categories and an "Other" folder that eventually holds half the company. 1.18 replaces that with a canonical registry of roughly 120 document types across 8 sections: corporate structure, governance, equity instruments (SAFEs, side letters, ROFR/co-sale agreements, warrants, vesting agreements, and more), team, IP, compliance & KYC, material agreements, plus a new Financials section for statements, audits, budgets, and bank letters.
Why ~120? Because that's roughly the shape of a real diligence request list. When an investor's counsel asks for "all agreements granting rights of first refusal," a taxonomy that only knows "Legal - Other" can't answer. A registry that distinguishes a ROFR/co-sale agreement from a side letter from a voting agreement can.
The registry is reachable through one searchable, grouped type picker shared by the Edit dialog, the upload review list, and manual add. Pick a type anywhere and the document re-files into the right Data Room section instantly. Aliases resolve too: type NDA, AoA, or SPA and the picker finds the canonical type without making you learn our naming.
Renewals chain into versions
Company documents recur. Licenses renew annually, policies get restated, leases get amended, and in most folders each renewal just piles up as another file with a year in its name.
Worked example. Your Data Room holds a Commercial License 2024. You drop in the Commercial License 2025. Vquity 1.18 detects it as a renewal of the existing document, chains it to the same series with an incremented version number, and shows a "v2 of Commercial License" badge right in the upload review. The section table keeps showing one row, the current version, with 2024 one chevron away in the version history.
This matters most in diligence, where the question is never "do you have a license?" but "is this license current, and can I see the prior one?" A versioned series answers both in one row; a folder of look-alike PDFs answers neither.
Extracted fields you can fix, with an audit trail
AI extraction is genuinely useful and occasionally wrong, and a system that doesn't admit the second half of that sentence isn't safe to keep records in. Since 1.10 the Data Room has extracted key fields from uploaded documents; as of 1.18, every extracted value in the document preview is editable in place. Click the field, correct it, done. When the field maps to a real cap-table column, the correction writes back to the cap table too, not just the extraction record.
Worked example. You upload a post-money SAFE. The extractor reads the purchase amount, $250,000, correctly and confidently, but pulls the valuation cap as $4,000,000 from a superseded schedule when the executed version says $5,000,000. In 1.17 that meant a round-trip through an admin screen. In 1.18 you click the cap field in the preview, type the correct value, and both the extraction record and the SAFE's cap-table entry now read $5,000,000.
Every correction is recorded: before and after values, who changed it, when, and the AI's original confidence for that field. Corrected fields wear an "edited" chip so reviewers can see at a glance which values a human has confirmed. The audit trail means an edited field is more trustworthy than an untouched one, the opposite of silent overwrites in a spreadsheet.
The AI Extractions center
The old admin "Extraction Debug" surface (built for us, tolerated by everyone else) has been rebuilt as AI Extractions, a proper working view. Headline stats up top: extractions run, source files, fields captured, low-confidence values, and manual edits. Below, a searchable activity feed of every extraction, filterable per document kind and by "Needs review," with expandable field grids showing each value's confidence and its override history. The Coverage and Schema panels carry over for anyone who wants to see exactly which fields the extractor captures per document type.
The workflow we designed for is triage: filter to "Needs review," walk the low-confidence values, confirm or correct each one in a couple of clicks, and watch the low-confidence count fall. It turns extraction quality from something you hope about into something you finish.
Readiness score, zip download, and table polish
1.18 also adds a diligence readiness score: a 12-item investor checklist (incorporation, constitution, licenses, statutory registers, resolutions, cap table, equity instruments, ESOP, founder agreements, IP, KYC, and financials) scored against what's actually in your Data Room, with an expandable what's-missing view.
Worked example. A seed-stage company uploads its closing folder and scores 8 of 12: incorporation, constitution, resolutions, cap table, instruments, ESOP, IP, and licenses check out, while statutory registers, founder agreements, KYC, and financials are missing. That's a to-do list you can clear in an afternoon, months before anyone sends a request list.
Smaller but daily-life improvements shipped alongside it. Select any set of documents and download them as a single zip. When someone asks for the corporate folder, you send one archive. Crisp per-format file icons (PDF, Word, Excel, images) replace the old emoji, and all section tables now share a fixed column layout, so rows align no matter how long a title runs. The Edit Details dialog was redesigned into a spacious two-column layout: title, type, description, and tags on the left; a segmented status control, executed and expiry dates, searchable linked-record pickers, and file facts on the right.
One fix worth naming
Clicking Settings, or following any settings deep link, from inside the Learning Center now navigates correctly instead of silently snapping back to Learn. Small bug, disproportionately annoying; fixed.
Why this release, and what's next
Most startups assemble their data room in a two-week panic after the term sheet arrives, and diligence goes exactly as well as that sounds. Our position is that the data room should be a living part of the company record, filed and scored continuously, so that "send us your data room" is a link, not a project. If you want the independent case for that, building a diligence-ready data room covers what goes where, and the academy lesson on due diligence data rooms fits it into the closing process. Everything in this post ships in Vquity's Data Room, included in the one all-inclusive plan like every other module.
What's next: extraction coverage keeps widening across the new taxonomy (the Coverage panel in AI Extractions shows exactly where it stands at any moment), and the readiness checklist will keep sharpening as more real diligence lists pass through it. Tell us what your investors asked for that the score didn't catch.
The full change list for this and every prior release is in the changelog. To see the new Data Room on realistic data, open the app at lisan.org/vquity and load the seeded sample company, or grab the Windows desktop app, same data, native window.