LiveDrop the deck. Draft the memo.
Deal Memo AI turns a pitch deck and a data dump into a structured, cited investment memo. Every figure is pulled deterministically by parsers — the model only narrates.
Why
Diligence means assembling the same memo by hand every time — copy a number off a slide, cross-check a spreadsheet, paste it into a template. It takes a day, and the figures still need re-verifying.
How it works
Drop your files
Pitch deck, financials, founder emails — PDF, PPTX, XLSX, CSV or notes. A whole folder or a handful of files.
Facts get extracted
Parsers pull every number — revenue, ARR, burn, runway, raise, valuation, retention — and tag each with its source locator. No model touches the figures.
The memo drafts itself
Bring your own key. The model writes the prose around the facts, cites locators, flags gaps, and gives a Pass / Explore / Strong-interest read. Export as Markdown.
Features
Deterministic extraction first
Numbers come from parsers, not a model. Section 3 — Traction & financials — is rendered straight from the extracted facts and never passes through the LLM.
Every figure cites its source
Page, slide, sheet and cell. Each value links back to the artefact it came from, so the whole memo is auditable, not just plausible.
Bring your own key
Anthropic, OpenAI, Google or Groq. Your key stays in your browser and is sent per request — never stored, no server-side secrets.
Honest about gaps
Missing data reads “Not found in provided materials” — never fabricated. Scanned PDFs are flagged as needing OCR, not silently dropped.
Fixed 8-section template
Snapshot, traction, market, team, terms, strengths & risks, recommendation. Consistent and comparable across every deal you run.
CLI or local web UI
A Python CLI for the terminal and a drag-and-drop web app. Runs on your machine; nothing leaves it except the model call you authorise.
Who it's for
Built for operators, angels and analysts doing first-pass diligence — the memo you'd otherwise spend a day assembling.
Open & honest
FAQ