Everything here is running, not rendered.
Five working apps, in your browser, no login and no sandbox data. Bring a real document — a print, a bid sheet, a vendor quote — and watch it come apart into fields you can edit.
These are the same tools we install, not marketing builds of them. So they will occasionally flag a line they were unsure about rather than pretending to certainty — that is the behaviour, not a rough edge.
The quoting workbench.
Three readers on one workbench, sharing a pricing library and a history. They differ in what lands on the desk: a customer's print, a district's bid sheet, or a vendor's quote. Each one is live — bring a real document rather than a sample, since reading a messy real page is the entire point.
Quote Workbench
- Bring
- A print or drawing (PDF, PNG, or iPhone HEIC)
- Takes
- ~2 min
Drop in a customer print — PDF, drawing, or a photo off the shop floor. It extracts dimensions, material, tolerances and finishing, pulls the bill of material apart line by line, tags each one fabricate / sourced / hardware, and prices it against your prior-quote library.
Open Quote WorkbenchBid Sheet Capture
- Bring
- A marked-up EDS bid sheet
- Takes
- ~90 sec for 7 pages
Drop in a district bid sheet with your pricing and compliance notes handwritten on it. Every line comes back as data — the printed spec and your written response in separate fields, pack sizes intact, no-bids preserved, extended price computed in software, and the uncertain lines flagged.
Open Bid Sheet CaptureVendor Quote Capture
- Bring
- A vendor quote or price sheet
- Takes
- ~1 min
The mirror image: a price a supplier sent you, as a PDF, a forwarded email thread, or a photo of a marked-up quote. It splits the vendor from the contact, pulls every line item with its SKU and price, and reads the pricing basis — the /M and /C traps that packaging suppliers use, where mistaking per-thousand for per-each is a 1,000× error. Out as JSON or CSV in Zoho Quotes field names.
Open Vendor Quote CaptureAgents and models.
These two need nothing from you but a URL and a few minutes. The agent builder stands up a working sales agent for your own site; the leak model is the diagnostic we use to work out which of these apps is worth your money in the first place.
Build Your AI Sales Agent
- Bring
- Just your website URL
- Takes
- ~3 min
Paste your website URL. It reads your site, builds a sales agent tuned to your catalog and your buyers, and drops you straight into a live conversation with it — the midnight-inquiry, spec-question, route-to-the-right-rep experience your buyers would actually get.
Open Build Your AI Sales AgentRevenue Leak Model
- Bring
- Your own rough numbers, or use the defaults
- Takes
- ~4 min
An interactive model of where demand drains out between the first click and the paid invoice — capture, quoting, follow-up, fulfillment. Watch the RFQs spill at each stage, then pick three fixes and see how much margin comes back.
Open Revenue Leak Modelyou start
Nothing you upload leaves your browser except to be read. The workbench keeps your library and history in local storage on this device — so it is private to you, and it also means a different browser or machine starts empty.
The pricing library is what makes quoting good. On a fresh browser there is nothing to match against yet. Load a handful of prior quotes in the pricing library first and the matcher has something to work with.
Flagged lines are the feature. Each reader tells you which lines it was unsure about instead of presenting one confident answer for everything. In a real install that is where a person spends their review time.
Want it pointed at your own documents?
Bring three real documents to a 20-minute call — a print, a bid sheet, whatever your team retypes most — and we will run them live rather than talk about it.