The engine that reads a software business — and prices it.
Caliber connects read-only to a micro-SaaS company’s revenue, repositories and analytics, lets code evaluate code and models evaluate the financials, and returns a calibrated value band, a quality grade and a real, revisable offer — in an afternoon, not a quarter. Built to price fairly, show its work, and pass on what it can’t stand behind.
Connect. Measure. Value. Offer — or pass.
The bottleneck in buying a small software company isn’t judgment — it’s the weeks of manual diligence that cost the same whether the business is worth $300k or $300M. Caliber turns that read into an afternoon by evaluating the primary sources directly.
Connect
Read-only access to billing and subscriptions, the code repositories and product analytics. Nothing is written, nothing leaves without consent, and the founder sees exactly what was read.
Measure
Code evaluates the code — architecture, coverage, dependency rot, key-person risk. Models evaluate the money — revenue quality, retention, CAC payback, margin, concentration. Each with a confidence.
Value
A calibrated intrinsic-value band and a quality grade — not a single false number — with the weakest axis flagged.
Offer — or pass
A real, revisable offer the founder can take to their own advisor — or an honest Pass, with reasons. Never auto-executed; a human is accountable for every offer.
A grade and a band, because one number hides the risk.
A single sticker price averages away the thing that kills a deal — a business can be profitable and still un-underwritable. Caliber reads the axes that fail on their own and shows which one caps the value, weighting to the weakest link, not the mean.
Revenue quality
Retention, churn shape and the gap between booked and collected.
Unit economics
CAC payback, margin, and profit that depends on paid acquisition.
Code health
Architecture, coverage, dependency rot and bus-factor — read by code.
Concentration
Customer, channel, platform and key-person dependence.
CI ±0.2 · priced off the weakest axis
The most important output is a calibrated “Pass.”
An engine that always finds a reason to buy is just a lowball machine with good manners. Caliber is built to decline — a single customer that is 70% of revenue, a codebase no one can safely change, retention that looks fine until you age the cohorts. Pass is a first-class result, returned with its reasons, and pass-rate is a published quality metric, never overridden because there is capital to deploy.
Pass — a first-class outcomeBelow the line human M&A won’t cross, software is illiquid. Caliber makes it liquid.
Manual diligence costs roughly the same at $300k and $300M, so beneath about $5M the deal economics collapse — nobody can work the deal for a fee it can bear. The result is a vast long tail of profitable micro-SaaS that can’t be bought or sold at a fair price. Drive the marginal cost of a rigorous read toward zero and that long tail becomes a liquid, underwritable asset class — priced on fundamentals, held or improved, disciplined by a published loss history. Liquid is not frantic: this is hold-or-improve underwriting, not a high-frequency flip.
Four things compound. The language model is none of them.
A competitor can call the same models. What Caliber accumulates instead is harder to copy and grows with every deal it touches.
Published value-vs-outcome curves, weakest-link pricing and pass discipline.
A proprietary dataset of offers, passes and what the businesses did next.
Repeatable, humane turnaround and infrastructure engineering — not a prompt.
A reputation for fair offers, built to the rules on the horizon rather than around them.
The clean exit the long tail never had.
Today a founder with a healthy $400k-ARR product faces months of broker back-and-forth, stacks of NDAs, a tyre-kicking buyer who ghosts, and a lowball at the end. Caliber is built to be the opposite: a rigorous read in an afternoon, the whole model shown to you, a fair price you can walk away from, and a buyer that keeps your product and your customers alive.
The old way
- Months of broker back-and-forth
- Stacks of NDAs, a tyre-kicker
- A lowball at the end
With Caliber
- A rigorous read in an afternoon
- The whole model, shown to you
- A fair price you can walk from
An engine that buys companies needs brakes. These are the brakes.
The easy version of this is ugly: lowball the desperate, strip the costs, delete the people, flip the shell at a markup. That version is simple to build and impossible to defend — and it is the version we refuse. So the guardrails here aren’t fine print; they are the structure of the product.
No lowball ambush — the valuation and its full reasoning are shown before any offer, and tracked to real outcomes.
No strip-and-flip — value comes from removing real waste and growing, never from gutting a product to resell an empty shell.
No coercive lock-in — an offer is revisable and non-binding until the founder accepts, with their own advisor in the loop.
No treating people as line-items — acquired products, customers and teams are sustained, not deleted; changes are made in the open.
No gameable metrics — Caliber prices durable fundamentals and audits for vanity signals engineered only to trigger an acquisition.
No overriding a Pass with capital appetite, and no acquisition without a human accountable for the price and the plan.
The honest answers.
Isn’t this just asset-stripping and flipping with a nicer UI?
That’s the version we refuse to build. Caliber’s return comes from pricing fairly and then removing genuine waste and growing a product — not from gutting a company to resell an empty shell. It is built to hold or improve, the valuation and its reasoning are shown to the founder in full before any offer, and a Pass overrides capital appetite. Strip-and-flip is a named red line, not a growth tactic.
What happens to the people — the analysts and lawyers, and the team you acquire?
Honestly: cheaper diligence moves the work that bankers and lawyers do on sub-$5M deals, the same way cheaper anything reshapes the work around it — and pretending otherwise would be dishonest. What Caliber is built not to do is treat that as a licence to delete people for sport. The cost it removes is diligence friction; acquired teams, customers and products are things to sustain, not line-items to cut, and any changes are made in the open with a human accountable.
A binding offer in seconds sounds coercive. Is it?
It would be, so that isn’t what this is. Caliber can produce a fast offer, but the offer is revisable and non-binding until the founder accepts, their own advisor is in the loop, and nothing auto-executes — a named human is accountable for every price. Speed is for the founder’s benefit, not a pressure tactic.
How do you avoid lowballing founders who need to sell?
Transparency and calibration. The founder sees exactly how the number was built, axis by axis, and can take it to anyone. The engine is calibrated to a fair intrinsic value and tracked to what the businesses actually do next, so a pattern of lowballing would show up in the loss history as mispricing — the opposite of the reputation the whole model depends on.
Won’t founders just engineer their product to trigger an acquisition?
Some will try — that’s Goodhart’s law, and we treat it as a design input. Caliber prices durable fundamentals (retention aged by cohort, collected revenue, maintainable code) rather than gameable vanity signals, and audits for metrics that look engineered. Signals that don’t survive scrutiny lower confidence and push toward Pass, not toward a higher offer.
The business, measured before the money.
Caliber is an early concept from The Adjacent Possible. We’re building the read-only diligence engine, the calibrated grade-and-band, and the loss-history ledger — and we’d like to pressure-test the guardrails with founders weighing an exit, operators who buy and run software, and capital partners who want a disciplined, fair way into the long tail.
Or write to [email protected]
Caliber is not a broker, a bank or a registered investment adviser, and offers no regulated financial, investment or legal advice.