Is our pricing consistent?
Finds every figure across the set and flags it when two of them disagree.
A working demo, not a mockup
Pricing notes, call notes, positioning memos, competitor research. Load them in, ask a question the way you would ask a coworker, and every answer shows the source text underneath it.
No login. Runs in your browser. Load the demo set in one click or bring your own files.
Is our pricing consistent?
No. Two figures show up across the set. Starter is listed at $999 a month, and the sales notes describe a $25k annual pilot for enterprise banks. 1
Worth a look before either number goes in front of a customer.
Pricing Summary
The current pricing workstream contains an unresolved tension. The early packaging model includes a $999/month Starter plan for mid-market fintech teams, but recent sales notes suggest enterprise banks expect a scoped annual pilot with a $25k minimum.
Every revenue team writes the facts down somewhere. A pricing doc from March. Call notes from Tuesday. A positioning memo the founder wrote that three people read. When a rep needs the answer in the next ten minutes, they ask in Slack, guess, or reuse a deck from last quarter.
Markdown or text files, or paste text straight in. Your files stay in the browser. There is no account and nothing to set up.
Each passage keeps its file name, its section heading and its line numbers. That is what makes an answer traceable later.
It finds the passages that match your question, writes the answer from those, and numbers every one it used.
Click a number. The source document opens beside the answer with the exact lines marked. If the wording is off, you see it in a second.
Is our pricing consistent?
Finds every figure across the set and flags it when two of them disagree.
Which claims would not survive review?
Pulls the absolute wording. Guaranteed, fully automated, zero risk.
Who are we selling to?
Shows you where the memo and the call notes describe a different buyer.
What do buyers push back on?
Reads the objection notes and gives you the lines back with sources attached.
Documents are split at their headings. Each passage is scored against the question using term frequency weighted by how rare each word is across the set, with a bonus when the words land in a heading. The top passages come back carrying their line numbers, which is how the reader knows what to highlight.
Set an OpenAI key and those same passages get sent out for the write-up. Leave it unset and the answer is assembled from the passages directly. The tool labels which one you are looking at, every time.
Next.js App Router, React 19, TypeScript, plain CSS, GSAP and Lenis for the motion on this page.
Accounts, a real vector store with embeddings, permissions, connectors to Drive and Notion, re-indexing in the background, and scoring against a fixed question set.
Seven documents from a made-up software company, written to disagree with each other on purpose. Find the contradictions in about two minutes.
Open the tool