SMALL SOFTWARE · HARD EVIDENCE

Focused workflow software, tested in the open.

We build narrow experiments for expensive operational problems—and let customer evidence decide what deserves to exist.

Explore the experiments ↓Start a conversation ↓

LIVE VALIDATION TESTS

11 narrow bets. Evidence decides what survives.

Each page states what is real, what is still a hypothesis, and the commitment required for the idea to advance.

Assignment Guard

Explain early-assignment risk before it becomes a support ticket.

View experiment →02

Lotline

Find cross-account tax-lot exceptions without pretending tax is simple.

View experiment →03

CostSplit

Stop eating unbilled API and cloud fees.

View experiment →04

PermitSync

Know when a permit changes before a crew reaches the site.

View experiment →05

SourceDraft

Finish customer compliance questionnaires without hunting through folders and old spreadsheets.

View experiment →06

AgentScope

Know what every AI agent can read, change, send, or delete.

View experiment →07

MCP Preflight

Review your MCP server before customers connect real credentials.

View experiment →08

EvidenceKit AI

Answer AI governance questions with evidence, not a last-minute policy scramble.

View experiment →09

MRF Preflight

Check a hospital price-transparency file before an executive attests to it.

View experiment →10

DirectoryDiff

See where payer directories disagree about your practice.

View experiment →11

Roster Preflight

Catch provider-roster errors before a payer sends the spreadsheet back.

View experiment →

HOW WE DECIDE

Thirty days. Qualified evidence. A visible decision.

An experiment continues only when the right buyers demonstrate a painful current workaround and make a concrete commercial commitment.

100+qualified visitors

5+buyer conversations

3+pilot requests

1–2paid commitments

Revise when qualified buyers engage but the offer does not convert. Stop when targeted distribution produces no repeated problem evidence. A paid commitment means collected payment, a signed pilot, or a written commitment with price and start date.

A SPECIFIC PROBLEM?

Start with context, not a calendar link.

Tell us what is expensive, repetitive, or difficult to evidence. If it matches an active experiment—or deserves a new one—we’ll respond personally.