The PPE1 Platform

Everything a lender needs to price, lock and govern its loans.

PPE1 is a complete Program & Pricing Engine platform: six products and two platform features, sharing one database, one sign-in and one pair of engines. It was designed by our founder, whose 40+ years in mortgage banking include 20 running secondary marketing and 34 building mortgage technology, much of it at the same time, and it was built over four years by our founder and a team of senior developers. One lender or service provider can own it outright, or many companies can license its full source code.

310K+
lines of hand-written code
530+
stored procedures behind a governed data layer
200+
database tables, with 114 foreign keys, every one trusted
51
system-versioned tables keep full row history, so a loan can be priced as of any moment
7,400+
automated tests across five solutions
<2s
per loan for full best-execution pricing
<0.2s
per loan when a whole bid tape is priced
40+ yrs
of mortgage banking experience behind the platform

Measured on the platform itself, October 2026. Public figures are rounded down.

One platform, not a bundle

Every product below was built for this platform. None was bought and bolted on, so they share the same data, the same security and the same answers.

One database
Defined once, priced the same way

Manager and the engines read and write one database through stored procedures. The program an administrator defines in Manager is the program the engine prices. There is no sync job and no second copy to drift.

One sign-in
One security boundary

A single sign-in carries each user's roles to every product: on the web, on the phone and through the API. Two-factor authentication, sessions and access rules are enforced in one place.

One pair of engines
The same price wherever it is asked for

A loan officer's quote, a mobile quote, a 5,000-loan bid tape and a loan sent from your LOS all run through the same Program Engine and Pricing Engine. A price does not depend on which door the loan came in by.

Explore the platform

Choose a product or a platform feature. Its story opens right here, and Read more takes you as deep as you want to go.

Products
Platform Features
Product · The control plane

Manager

Everything the engines use is defined, scoped and governed here.

Manager is where a lender sets up the business the engines price for: client companies and their org structure, investors and their programs, eligibility guidelines and price adjustments, the ratesheets that become base pricing, and the government data behind eligibility.

Nothing reaches a loan until it has been defined here by someone with the right role, and every change is kept. Its four modules follow the way a secondary marketing department actually works, from onboarding a new investor to loading this morning's rate sheets.

4
modules: Company, Program, Ratesheet and Utility
>1M
ratesheet pricing records migrated per day
Full
history on every configuration change
Twice
role checks: the screen hides, the server refuses

The four modules

  • Company Manager. Client companies, regions, branches and users; broker (TPO) divisions; investors and their programs; and program subscriptions, the switch that decides what a client may ever be priced on.
  • Program Manager. Guideline matrices for credit, LTV and CLTV, loan amount, DTI, reserves and MI; overlays; LLPA and SRP grids; and inheritance, so common rules are entered once and reused.
  • Ratesheet Manager. Maps each investor's rate sheet layout (block or linear) to products, migrates it into base pricing, and holds each program's lock and extension policy.
  • Utility Manager. FHFA and HUD loan limits, census tracts, AMI and Duty to Serve data from the official publishers, Federal Reserve index and rate series, and the lock desk's Rate Lock Directory.

How a program goes live

  1. Define the investor and the program, with its channels and price files.
  2. Build the guidelines and overlays that decide eligibility, and the LLPA and SRP grids that adjust price.
  3. Map and migrate the investor's rate sheet. Show-Pricing lets the desk check the result before any loan sees it.
  4. Subscribe the client companies that may sell the program. The engine considers it from the next quote.

Under the hood

Access
Named roles on every screen and action, scoped to the regions and branches a manager runs
History
51 system-versioned tables keep every prior version of a row; nothing is hard-deleted
Data layer
Stored procedures through Dapper; inline SQL limited to a few reviewed exceptions
Imports
Government data loads in two steps: stage and preview, then apply

Why it matters

A pricing engine is only as good as the configuration behind it. Manager makes that configuration explicit, owned and reviewable, so a lender can say exactly why a program priced the way it did on any given day, and who set it up that way.

One deployment can host many independent lending companies side by side, each with its own hierarchy and its own broker division, every record scoped to its owner.

Product · The engines

Loan Pricer

Does this loan qualify, and at what price? Answered in under two seconds.

Loan Pricer holds PPE1's two engines. The Program Engine works out which program and product combinations a loan is eligible for. The Pricing Engine then builds the final rate and price for each one and ranks them by true best execution.

The two engines stay separate on purpose. Eligibility can be audited on its own, and priced results can be ranked again without re-running the guideline logic. The same engines answer a loan officer's quote, a mobile quote, a bid tape of thousands of loans and a loan sent from an LOS.

<2s
per loan, full-scope best execution
6
eligibility stages, always in the same order
Any date
a loan can be priced as of any past moment
Field-level
audit of why each program passed or failed

A loan's journey

  1. Submit. A guided form offers only the programs the client is subscribed to, and geocodes the address to census tract, county, AMI and Duty to Serve.
  2. Eligibility. Six stages narrow every subscribed program to the program and product pairs that qualify. An overlay can only tighten a guideline.
  3. Pricing. Base price, LLPA and SRP adjustments, ratesheet overlays and the lock period, then a best-execution ranking.
  4. True BestEx. Rates down the side, products across the top, the best price in each cell. Every cell opens to its full breakdown.
  5. Counter-offer. When nothing fits as entered, the engine relaxes one constraint and comes back with the best terms actually available.
  6. Lock. The officer requests a lock from the grid, and the lock desk takes it from there.

What sets it apart

  • Priced as of any moment. Every engine entry point accepts an as-of date, so a locked loan re-prices against the rules and rate sheet of the day it was locked.
  • The code or the data? Loan Audit replays eligibility from the saved run, field by field, and names the guideline or overlay behind each result, even for a loan priced months ago.
  • The whole instrument. ARM margins, rate caps and prepayment terms are resolved along with the price.
  • Configured, not coded. Guidelines and overlays are authored as data, so a lender changes how the engines behave without a deployment.
  • Credit-aware. For Non-QM loans, graded credit events filter eligibility before pricing runs.

Under the hood

Program Engine
A chain of six stages (setup, header, finance, product, overlay, outcome), each passing an accumulating scenario to the next
Pricing Engine
Base pricing, adjustments, overlays and lock resolution, then ranking
Run modes
The full rate matrix for an interactive quote; the note rate alone for batch work
Data
The rate sheet grid is the platform's largest dataset; the engine reads only the rows that can apply to the loan in front of it
Proprietary
the adjustment math and the ranking, shown in the private data room
Product · Institutional scale

Bulk Pricer

A 250-loan bid tape priced in under 49 seconds, without touching a live loan.

Secondary marketing works in tapes, not single loans. Bulk Pricer takes a spreadsheet of loans, maps its columns once, prices every loan through the same engines a loan officer uses, and turns the result into a dashboard and an investor-ready bid.

It runs in its own workspace. Bulk loans live in per-client staging tables and are priced in a no-write mode, so a desk can price, re-price and throw away thousands of loans with no risk to the production pipeline. Mark-to-market, a bulk bid and a what-if before a pricing change all start here.

<49s
for a 250-loan batch
<0.2s
per loan, priced in parallel
1000s
of loans in a single batch
0
writes to the live loan tables

From spreadsheet to bid

  1. Upload a bid tape, a portfolio extract or a what-if set.
  2. Map columns and coded values to loan fields once. The map is saved and reused for every tape in that format.
  3. Process. Test one row's eligibility, price one loan in full, or run the whole set as a tracked, parallel job.
  4. Review the dashboard: priced, failed, counter-offer and no-pricing, with weighted-average coupon, price, FICO and LTV.
  5. Export a Bid Summary PDF grouped by plan type, each group carrying its own weighted averages.

Built for volume

  • One engine. Each row becomes the same loan object the single-loan form builds, so a price on a tape is the price that loan would get on its own.
  • Fast by design. Batch loans price at the note rate only, in parallel, with a fresh engine chain for each loan.
  • No message broker to run. Large uploads queue in-process and price on a background worker while the screen shows progress.
  • One bad row never stops a run. Rows with missing or unmappable values are reported against their spreadsheet line and skipped.
  • Safe to re-run. Reprocessing against today's pricing can be cancelled, and cancelling restores the prior results.
Product · In the field

Loan Pricer Mobile

A quote on the phone is the quote the web gives, because it is the same engine.

Loan officers price at open houses, in cars and across kitchen tables. Loan Pricer Mobile gives them the full pricing experience on a phone: a guided entry wizard, true best execution in swipeable cards, counter-offers and the loan audit.

Web and mobile are two front ends over one shared core, so the loan model, the services and the validation rules are written once. A rule change reaches both on the same day. The phone signs in through the same identity service, with two-factor authentication and Face ID.

1
shared core for web and mobile
0
copies of the pricing logic
iOS & Android
native app; store release underway
Face ID
sign-in on the platform's identity

On the phone

  • Price a loan through a guided, validated wizard. Only the client's own programs appear.
  • True BestEx as swipeable cards sized for a thumb.
  • Counter-offer with ranked, achievable terms when a loan does not fit as entered.
  • Loan Audit with pass or fail by program and the field-by-field reason, so an officer can troubleshoot anywhere.

Secured like the rest of the platform

  • The same account, sign-in and two-factor flow as the web.
  • Sessions refresh quietly in the background and log out after idle time, so a phone left in a pocket is not an open door.
  • Each device carries its own refresh token, so signing in on one never signs another out.

Why one core matters

A separate mobile calculator drifts away from the real engine over time, and the officer in the field ends up quoting a number the lock desk cannot honor. PPE1 has no separate calculator. The number an officer quotes on the phone is the number the desk sees.

Product · Guided by eLCee

Learning Center

Answers drawn from the platform's own documentation, at the depth each person needs.

The Learning Center replaces the 400-page manual nobody reads. Its guide, eLCee, answers questions about any part of the platform, from setting up a program subscription to how the eligibility engine is built, and shapes each answer for a business, operational or technical reader.

Its answers come from a curated library of more than 110 knowledge documents, written against the live code and checked again whenever the code changes. An evaluation suite tests every change to that library, and access tiers decide how deep each reader can go.

110+
curated knowledge documents
5
solutions covered end to end
3
audiences: business, operational, technical
Tiered
answers scoped to the reader's access

Ways in

  • The book. Written summaries of every area, layered by depth.
  • Ask live. A conversation with eLCee that remembers what was asked before.
  • Diagrams. Architecture for the technical reader, workflow charts for the operational one.
  • Walkthroughs. Step-by-step help for any task.
  • Code on screen. For technical evaluators, allowed code shown and explained, behind an access gate.

How it stays right

  • eLCee reasons over the platform's own documents, never the open internet.
  • What she reveals depends on the reader's access tier, and a hard guard keeps secrets and credentials out of every answer.
  • When a question exposes a gap, the library is strengthened, so answers improve over time.
  • Automated evaluation and adversarial tests run on every change.

For a buyer or licensee

You do not have to take our word for how the platform behaves. Ask eLCee how eligibility, pricing, overlays or history work, and the answer comes from the documented behavior of the running system. Your staff get day-one productivity; your engineers get the confidence to own it.

Product · Security

Identity & Security

One security boundary for the web, the phone and the API.

A dedicated authentication service issues one signed token that Manager, Loan Pricer, Bulk Pricer, the mobile app and the API all trust. Users sign in once and move across the platform with their roles intact.

Sign-in, two-factor, refresh and the account lifecycle all live in one service, so a security reviewer has one well-defined surface to assess. Every sign-in, sign-out and timeout is written to an access audit trail.

5
surfaces on one identity
2FA
the same on web and mobile
0
server-side stores for two-factor state
1
security boundary to audit

The security model

  • Single sign-on. One account spans the platform; every product and the API honor the token issued at login.
  • Stateless two-factor. A short-lived signed token carries the "verified, awaiting code" state, so it works the same on web and mobile. Codes are hashed at rest.
  • Per-session refresh. Each session rotates only its own refresh token, so devices never sign each other out.
  • Claims carry capability. The token carries the user's roles, so each service enforces access from what it already holds.

Hardened by default

  • No account enumeration. Sign-in, recovery and registration replies never confirm whether an account exists.
  • Single-use tokens. Reset and activation links are random, stored only as a hash, time-boxed and consumed on first use.
  • Rate limits on sign-in, recovery and two-factor resend.
  • Anti-phishing recovery. A user must pick their real company from decoys before any reset is sent.
Platform feature · Built in

LOS Integration

Built to connect to your loan origination system.

PPE1 includes a documented LOS integration API. Your LOS sends a loan; PPE1 checks it is ready to price, runs it through the same engine your users already trust, and keeps a synchronized copy that your loan officers open in place from the LOS, signed in as themselves.

Because you license PPE1 and run it in your own environment, the connection is yours too: your LOS talks to your PPE1 through your own API access, and your loan data never leaves your control. The integration is LOS-neutral. It is designed for any LOS that lets your organization read and update loans through its API.

Documented
LOS integration API
Mapped
fields and coded values, by configuration
Same
engines your loan officers already use
Yours
your environment, your API access, your data

How a loan arrives

  1. Your LOS sends the loan to your PPE1 through the integration API.
  2. PPE1 checks it is ready to price and answers with plain-language messages your team can act on.
  3. The engines price it, and PPE1 keeps a synchronized copy of the loan.
  4. Loan officers open it in place from the LOS, signed in as themselves.
  5. A change to a locked loan sent by the LOS goes to the lock desk as a change request, priced as of the lock.

What your IT team owns

  • Field mapping. Your LOS fields mapped to PPE1's loan fields, through configuration.
  • Value mapping. Your LOS's coded values translated to PPE1's, so a "2" in one system means the same thing in the other.
  • No custom code inside PPE1. The mapping is data, so it survives every upgrade.
  • Validation you can act on. Every send is checked, and every problem is named in words your team understands.

The Lock Ledger, inside the loan

When a locked loan is opened from the LOS, its users see a read-only Lock Ledger tab: the full history of the lock and its pricing, measured the way the loan officer sees it. Decisions stay with the lock desk, and the desk's own figures stay with the people entitled to them.

Platform feature · Secondary marketing

Lock Desk & Lock Ledger

Loans change after they lock. Every version, request and adjustment stays on the record.

After a loan locks, it keeps changing: a new appraisal, a different loan amount, a request for more time, a concession to keep the borrower. PPE1 treats each lock as a commitment with a full history, run by a lock desk with the controls a secondary marketing department needs.

A loan officer can see a what-if price on a locked loan but cannot change the lock. What the officer can do is send a request. The desk decides it in the Rate Lock Directory without re-running the engine, because the request carries the values being decided, and every decision becomes part of the lock's Ledger.

Every
version and request kept in the Ledger
4 tiers
of compensation: Company, Region, Branch, User
Server
enforced visibility of desk-only figures
As of
the lock date: changes priced against that day's rules

What the Ledger shows

  • The full chronology, newest first: the lock, each revision, change and extension requests, exceptions, extensions, voids and restorations.
  • Who, when and at what price on every row, and every row opens to its complete pricing breakdown.
  • Desk-only figures (the corporate price, profit, rounding and the investor's expiration) only for users who hold the administrator pricing role. The server enforces it, not the screen.
  • The same Ledger from the LOS, read-only, for the loan's own users.

The desk's controls

  • Confirm the lock once the investor acknowledges it; the loan officer receives the confirmation by email.
  • Approve or reject change requests from the loan officer or the LOS. A change that does not affect the locked price is applied at once.
  • Extend three ways: an investor grant, the lock's own unused coverage, or a fee set by the program's extension policy. An extension can be voided to return the lock to its prior terms.
  • Grant pricing exceptions for a courtesy or an adverse event.
  • Void and restore. A voided lock is retired, never deleted.
  • See what needs attention at a glance, including every lock whose confirmation is owed.

Four-tier compensation that reconciles

Every lock carries its compensation split across Company, Region, Branch and User. When the desk grants a pricing exception, PPE1 records how much moved, how much reached the borrower's price and which tiers bear the rest. The tier totals are derived from the stored pricing when the exception is written, never taken from what a screen sends, so they always agree with the Ledger.

Repricing after lock is routine, and it is where margin quietly leaks: a concession nobody recorded, a change approved without its price, an extension nobody charged for. The Ledger is built to keep each of those on the record, by someone who ran secondary marketing desks for twenty years.

See it running.

Qualified parties get private, hands-on access to the live platform, and can read the Forensic Code Review before they ask a single question.

Request Private Access