Build on the AI-native lending and servicing platform
Onboard borrowers, originate and service loans, run a governed AI workforce, and give borrowers a white-label self-service experience — through typed APIs that apply current-fact policy at each owning boundary.
LendEasy is an AI-native lending platform: the LendEasy Lending Core — a complete system of record for installment loans, BNPL, and merchant advances — with a servicing control plane that runs natively on it, a named AI agent workforce (Nova, Cookie, Docgest, and Arya), an operator workspace, and white-label borrower apps. As an additional deployment option, LendEasy's servicing workflows can also run on a core you already operate.
System architecture
Human agents and AI agents share workforce, policy, and audit, but their execution authority stays explicit. Human servicing calls use typed owning routes and exact permissions; workflow and AI automation use registered operations or typed effects. Borrower-contact and governed-operation boundaries invoke the compliance engine with current facts, while the lending core remains the system of record beneath the servicing plane.
The platform in four parts
Each part has an explicit API and ownership boundary. The runtime composes them into one operating system for lending and servicing, anchored on the LendEasy Lending Core — and each boundary is a documented, typed contract where systems you already run can integrate.
The system of record the platform runs on natively: installment loans, BNPL, merchant advances, schedules, double-entry accounting, statements, and a Reg Z Appendix J APR engine whose disclosures an auditor can reproduce. Our servicing workflows can also run on a core you already operate.
Named agents that work cases alongside your team — Nova on the phones, Cookie beside every case, Docgest on documents, Arya on the inbox. Same queues, same permissions, same audit trail, instant kill switch.
Cases, tasks, and smart queues, plus a compliance engine invoked by owning contact and governed-operation boundaries — each decision grounded in current facts and retained evidence.
The operator UI where agents and managers resolve work across 20+ case types — full borrower context, configurable workflows, and a manager command center.
See it end to end
The best introduction is one worked case: a Harbor delinquency handled by the full roster — case opened, Cookie's grounded brief, Nova's governed call, a recorded promise, and exam-ready evidence at the close. Follow the delinquency walkthrough →
A portfolio you can follow
Every calculation and servicing guide works from four named, fictional contracts. Follow the same balances from origination through payments, delinquency, schedule changes, refunds, disclosure, and closure—without having to reverse-engineer a new example on every page.
Typed owning operations, exact permissions, and registered effects — one governed path for every mutation.
Start now or plan ahead
The servicing plane runs natively on the LendEasy Lending Core; as an additional deployment option, our workflows can also run on your existing core. Check API availability for the exact runtime boundary of the current build.
LendEasy Lending Core + servicing
- Authenticate and set your environment
- Onboard a customer & payment instrument
- Originate and fund a loan
- Open servicing cases and let the plane run
LendEasy servicing on your existing core
- Authenticate and set your environment
- Bind your system of record via the published fact/action contract
- Stream ledger facts in; receive governed actions out
- Open servicing cases against your loans (HTTP core binding is a roadmap contract — see bring your own core)
How this documentation is organized
- Guides — concepts & how-to: get-started, core conventions, and product docs for the Lending Core and the Servicing Plane.
- API Reference — every exposed endpoint, grouped by resource, generated from the OpenAPI spec.
- Recipes — task-oriented, end-to-end workflows that stitch several endpoints together, with the compliance checks called out at each step.