Tearsheet exists because the gap between “skim a listing” and “pay for a Quality-of-Earnings audit” was where most bad deals were hiding.
Self-funded acquirers and ETA searchers look at 30-80 deals a month. They can't QofE all of them — a real one costs $15K-30K and takes weeks. So they skim. The skim is where deal killers hide: aspirational add-backs, owner-dependent revenue, top-customer concentration, the kind of detail that surfaces in week 3 of diligence and costs you the deposit.
Tearsheet's job is the part between skim and serious diligence. Pre-LOI acquisition memos in the time it takes to read a CIM. Honest about what they don't know. Reviewed before delivery by an AI reviewer that's been told to argue with the drafter, with automated quality gates added for paid Deal Desk Sprint work.
The Acquisition Memo pipeline runs Plan → Research → Ground → Draft → Verify. The Ground step builds an Evidence Ledger that separates provided facts, derived checks, model-prior assumptions, and unknowns before the final memo is written. Target Scorer and Industry Teardown use shorter agent-specific pipelines, but the same rule applies: label uncertainty instead of hiding it.
URLs you paste are scraped via Firecrawl so the analysis is grounded in actual listing text, not whatever the model guessed. PDFs of broker packages are extracted server-side.
Tearsheet is built by Garrett Gleave, solo. No team, no investors, no growth department. The whole product runs on Next.js, Supabase, Stripe, Anthropic, and Vercel. The self-serve product is priced at $49/month for 100 memos; Deal Desk Sprint work is a one-time quality-gated service.
Security-sensitive changes get separate code and product review before deployment. The product is improved by reading every rating + feedback users leave.
Real human at hello@usetearsheet.com. If you spot a memo that's wrong, please tell us — we read every reply.