New assessment / Start
Form filled 0%

Free, private due diligence that runs in your browser

Check whether the numbers reconcile

Founders send a deck, a model and an update. VCR2 rebuilds the revenue from the company’s own customer ledger, recomputes every headline metric, shows where the stated figures disagree with the arithmetic, stress-tests the runway, prices the deal, and turns every gap into the question to ask on the next call.

  • Nothing you enter leaves this browser
  • Every point of the score is explained
  • Same inputs, same result, with a run ID
Northwind Datafictional sample, Series A
Net revenue retention in the deck134%
Rebuilt from the customer ledger98.8%
35.2 points apart, in the company’s favour. Nine more headline figures in this deck disagree with the arithmetic.

Question for the foundersWhich cohort and which period does the deck’s NRR cover, and does it include customers who left?

60+metrics derived from raw figures
29consistency checks
9forensic tests on billing data
30risk and strength patterns

Revenue rebuilt from the ledger

Drop in revenue by customer by month. VCR2 builds the MRR bridge, true cohort retention, concentration and cohort heatmaps: two days of associate work, in a second.

Stated against computed

If the deck says 134% NRR and the cohorts say 99%, that gap is the finding. A binomial test tells you when every error happens to flatter the company.

Forensics on the billing data

Pre-raise booking spikes, one-off charges booked as recurring, duplicate customers, quarter-end pull-forward and a Benford test, each with its method and its limits.

Runway odds, not a runway number

A seeded Monte Carlo gives the chance of reaching next-stage metrics before the cash runs out, and how much capital that takes at 80% confidence.

Deal maths to the exit

Pool shuffle, SAFE and note conversion, a term-sheet review against market standard, VC Method and First Chicago returns, and a full liquidation waterfall.

Learns from your deals

Save runs and import comparables to see percentiles against your own deal flow, and recalibrate the healthy ranges on it.

Also:

VCR2 is a workflow tool. It does not give investment advice and its score is not a recommendation to buy or sell a security. Both examples are fictional companies with synthetic data.

Step 1 of 4

What are you looking at?

Stage sets the benchmark thresholds and the dimension weights. Business model selects which metrics are scored and which are suppressed.

Stage
Business model

Step 2 of 4

Give it the source material

The strongest evidence is the company’s own data. Structured files are parsed in your browser, nothing is sent anywhere. Narrative material can optionally be read by a model you choose, and every extracted value keeps the line it came from.

Structured data · parsed locally, no AI

Customer revenue ledger most valuable

Revenue by customer by month from the billing system. Wide (one column per month) or long (customer, month, amount). Unlocks cohort retention, the MRR bridge, concentration and the forensic tests.

Drop a .csv or Excel .xlsx here, or click to choose
…or paste it
template

Monthly financials

Month by month: revenue, cash balance, net burn or total expenses, headcount. Any subset works. Unlocks burn trend, default-alive and a calibrated runway simulation.

Drop a .csv or Excel .xlsx here, or click to choose
…or paste it
Narrative material · optional AI extraction
⚠
Extraction sends the pasted text to the model provider you chose in the sidebar. Everything else in VCR2 (the ledger, the derivations, the checks, the simulation and the score) runs in your browser and sends nothing.

Deck, memo, founder update or term sheet

Plain text. Slide notes, an email update, a data-room summary, the term sheet.

Financial summary

A P&L, ARR waterfall or metrics tab. CSV or tab-separated works.

Or drop a .csv / .xlsx / .txt / .md file here

Uses your own key and your own account. Nothing is billed to the site owner.

Computing

Recomputing headline metrics from primitives…

Rebuilding revenue from the ledger
Deriving metrics from primitives
Running consistency and forensic checks
Scoring against stage thresholds
Simulating runway paths
Pricing the deal and matching patterns
Building the question list
Drafting the memo narrative
Idle