Meridian is the LightLeapAI platform for building any enterprise application you can imagine — starting with a prompt.
A CIO types one sentence. Meridian proposes a complete data model — objects, fields, relationships, roles, and rules — you review and confirm, and the platform writes a live, isolated tenant you can use in minutes. From there, the same conversation is how you change forms, build reports, chain workflows across your data, connect outside systems, and publish a public site that turns a visitor into an account holder.
For years, campuses and enterprises rented their software back from big tech. Meridian is the alternative — a platform you actually own and control.
SIS, ERP, CRM, and point tools — a patchwork of subscriptions that don't talk to each other, lock your data inside someone else's roadmap, and make a simple question mean an export, a ticket, and a data team.
Every meaningful change on a legacy system is a vendor engagement, a consultant statement of work, or a multi-quarter upgrade. You rent behavior you cannot change.
Intelligence is walled inside each separate app and billed as an upgrade, instead of being the substrate every workflow runs on.
Priorities are set for the average customer, not for you — and your institution waits in line for the features it needs most.
Meridian hands you the layers underneath the software — and lets you compose exactly what you need, in plain English, on infrastructure you control.
Describe a role, a tab, a workflow, or a data rule in plain language. The agent plans it, you confirm, and the platform writes it — every change audited.
Database-per-tenant isolation, field-level roles, and a single audited write path. Routine edits save with an Undo; high-stakes changes confirm inline first.
Your modules, your roadmap, optionally your fork of the code. AI is the substrate — every workflow can think, not just the ones you upgraded.
Meridian gives you the whole stack — the builder, the model, the APIs, the administration, and the data surface — composable and agent-native from line one.
Describe what you need in plain language. Meridian proposes a complete data model — objects, fields, and relationships — that you review and confirm before anything is written.
Create and edit objects and fields directly — types, requirements, origin, and field-level access — without code and without a release cycle.
A no-code, three-pane designer publishes a login-free public site — hero, forms, record lists, galleries. A public form can be the sign-up: a submission becomes the visitor's own account and first record.
Describe the report in plain language — or drag fields onto the canvas. Joins, anti-joins, filters, groupings, and count thresholds run through the same row-scoped, field-RBAC records API your app runs on. Every export audited.
Describe an automation in plain language. Meridian lowers it into a governed workflow that can follow a relationship chain across your data, fan out one record per match, and update fields on related objects — all approved inline before it goes live.
Connect any two systems — through Meridian or directly between them. Author the plan in words, edit it in words, and test it on one real record before it runs. Secrets are KMS-encrypted, tenant-bound, and rotatable without ever seeing plaintext.
Every object gets versioned REST APIs with OpenAPI docs, plus a live MCP server — so an agent (or the Meridian ChatGPT App) can sign in with LightLeap and drive your governed data as tools. Reads live; writes park for approval; the permission clamp travels with every call.
Explore every object's versioned REST API with live OpenAPI docs in-app — so your application is integration-ready the moment it exists.
Query, join, and create, update, or delete records in natural language — or drop into SQL — natively in the browser. Every change is previewed, confirmed, and audited.
Omnia is the conversational front door to the whole platform. Extend, enhance, and customize your application by talking to it — no support, vendors, or consultants required.
Meridian doesn't stop at the back office. Publish a login-free public site — a directory, a catalog, an application, an event page — from the same governed tenant, with a no-code three-pane designer. The public form is the sign-up: a submission writes an audited record and best-effort issues a passwordless magic-link invite, so the visitor becomes an account holder who owns the record they just created.
After the audited record write, the submitter gets a passwordless invite to their own email — their account stays inactive until they click it from their own mailbox. Open lead-capture or roster-only, your choice, per form.
Only the fields you explicitly expose go public. Public record lists are row-scoped and field-RBAC-projected — classified data is dropped even if you try to publish it.
Design with a block library and live desktop/tablet/phone previews — or describe the change to Omnify and confirm it. Publish once; it's an installable app on any phone, no App Store, no fees.
Open Meridian and describe what you need. The same platform that runs a full Student Information System spins up a law-firm practice system, an HR and asset platform, or whatever you type — objects, multi-hop workflows, versioned APIs, an agent hub, a public front door, and a natural-language data surface included.
Land on the templates picker with no login required. Pick a starter — Community-College District, HR, Law Firm — or type one sentence. Meridian provisions an isolated, database-per-tenant application on AWS Aurora and hands you the working URL. Total time: about thirty seconds.
Tell the platform what you need in plain language. Omnia researches the domain and streams back a complete proposal — objects, fields, and relationships — with a confidence score. You review and confirm before a single thing is written. Live on AWS Bedrock (Claude Sonnet 4.5) by default, with an automatic deterministic fallback so a demo never dies with the wifi.
Meridian gives you a live object designer — objects, field types, origin (core or custom), whether a field is required, and exactly which roles can edit it. Change the model without code and without waiting for a release. When the shape of the business changes, the shape of the application changes with it — the same day.
Type what you want to see in plain language and Meridian's planner builds the report and runs it — or drag fields onto a palette yourself. Joins and anti-joins ('records with no related X') are validated against the same relationships that drive the runtime. Filters, groupings, count thresholds, and CSV export run through the same records API your app runs on — so a report can never see a row a person couldn't. Every export is audited.
The Workflow Studio takes a plain-language automation — 'when a bursar posts a refund over $500, notify the CFO' — and lowers it into a governed AutomationDef you confirm inline. It isn't limited to one hop: a workflow can follow a relationship chain (Subject → Course → Section), fan out one record per match across several fields at once, set a field on a related object two hops away, and stamp SYSDATE. You review the trigger, conditions, and effects; the runtime writes it. No consultant, no statement of work.
Author a connection to email, a CRM, an LMS, a service desk, a payment provider — through Meridian or directly between two outside systems. A full-screen spine workspace shows the integration as a flow you can read: when it runs, what it reads, how it maps, what it writes. Change the match field or a filter in plain language and test it on one real record before it runs. Secrets are KMS-encrypted, tenant-bound, and rotatable without ever seeing plaintext. The eight-week custom-integration engagement collapses to an inline confirm card.
Every object gets versioned REST APIs with live OpenAPI docs — and a live MCP server. An agent connects and discovers your governed data as tools: reads run live, writes park for approval. The Meridian ChatGPT App lets a user 'Sign in with LightLeap' (OAuth 2.1) and drive their own workspace from ChatGPT — resolved through the exact same permission clamp as every other path, so it can never exceed its owner's authority or reach another tenant. The guardrails travel with every call.
Data Studio runs live on AWS Aurora. Ask a question of your records in plain language — 'list all matters where the client name contains Atlas' — or drop into SQL for the nerds. Query, join, create, update, and delete, with every change previewed, confirmed, and audited. Semantic search on objects and records via AWS Bedrock embeddings, with a deterministic fallback so it works offline.
Omnia is the conversational front door to the whole platform. Describe a role, a tab, a workflow, or a data rule and the agent plans it, you confirm, and the platform writes it — every change audited. Extend, enhance, and customize without calling support, vendors, or consultants. The AI proposes; the human commits.
Tell Meridian what you want to own that you don't own today — the rest is engineering, and we have fifteen years of it. Provision a live tenant and change it by conversation from day one.
Meridian's flagship tenant is a full agent-native Student Information System for Prime-Meridian Community College District — a working demo you can drive today. Every persona runs against the same governed runtime; every write pauses on an inline confirm; every action is audited under the tenant. A built-in Course Schedule Optimizer even staffs sections across terms and explains, in plain language, why any section couldn't be staffed and the single cheapest fix.
Register for class, pay a bill, check financial aid, and see the degree audit — all on the same platform, in plain language.
Create sections, run degree audits against program requirements, manage holds, and roll grades — with the audit trail inline.
District receivables, ledger, payment plans, and audited refunds. Every high-stakes write pauses on an inline confirm card.
ISIR verification, packaging within COA, SAP evaluation, R2T4 — the whole regulated workflow, honestly audited.
My sections, roster, grade entry, and census certification. Raise an early alert on an at-risk student — it routes to their advisor.
Consolidated student 360, Student Education Plans, early-alert triage from Faculty, and hold clearance — the loop closes.
The Faculty → Advisor early-alert loop is a real end-to-end narrative — Faculty raises the alert, Advisor triages it, Student sees the result — all on the same tenant, all audited.
Meridian was built next to a full-fidelity Banner-to-Meridian migration critic — plus one each for PeopleSoft, Workday, and Anthology — that grades every release. If the runtime can't faithfully hold your legacy shape, we say so; we don't move a hut into a mansion and paint over the walls.
“Think of Meridian as a mansion you're moving into. Don't try to move a hut into a mansion. Re-imagine your living conditions in the mansion. And then move the furniture you think is worth moving from your hut.” — Kiran Kodithala, Meridian webinar, August 27, 2026
Meridian ships five governance surfaces you can inspect on any tenant — and we don't just assert they work, we run N2N Services itself (HR, asset inventory, SOC 2 evidence, sales pipeline) on the same runtime. Isolation is a physical database boundary, not a hopeful WHERE clause. Every AI write pauses on a human. Every write — AI or not — writes an audit row in the same transaction. And the platform remembers what it learned so the next release doesn't repeat a known scar.
Every release is graded against a 13-article public constitution — small ships, verified trust, AI autonomy with human guardrails, truth over marketing, multi-tenancy, heartbeat, security, teacher-student ledger, attestation congruence, reliability, interpretability, 10× impact, continuous learning.
docs/mandates/QUANTUM_MANDATE.md — 13 articlesThe user sees, on every AI-driven surface, whether the reply came from live AWS Bedrock (Claude) or the deterministic offline planner. Live-by-default; automatic degradation on any error. Trust is what you show, not what you claim.
AI Builder · Omnia · Workflows · AI RulesEvery tenant gets its own physical Postgres database on AWS Aurora. Not a tenant-id column and a hopeful WHERE clause — an actual CREATE DATABASE, an actual boundary, an actual name in the connection string. RLS inside each tenant DB is a supplement, never a substitute.
AWS Aurora — RDS IAMEvery mutation writes a row to an append-only audit table under the tenant, in the same transaction as the write itself, with the actor's real principal, the resource, the before-and-after, and a scrubbed reason. There is no code path that mutates data without that row.
meridian.activity_log — governed storeEvery confirm card renders the relevant lessons the platform already learned from prior scars — pulled from a retrieval seam over the platform's own postmortems, ADRs, and learning-register entries. The next change is smarter than the last one because the platform remembers.
meridian.wisdom_evaluation — ADR-0058None of the above is a marketing claim you can't verify — every pillar names the file, table, or ADR you can read.
These are the LightLeapAI product line's numbers, not Meridian's — Meridian is new. They matter because Meridian is built on the exact same production stack (fraud intelligence, document intelligence, and the Helios integration spine) that already earned them. Platform figures as of Fall 2025, across the LightLeapAI product line; fraud and aid-protection totals are cumulative since launch.
Light Leap AI completely transformed our admissions process for Rising Scholars. What once took weeks of manual paperwork is now efficient, accurate, and lets our staff focus on what matters most. We’re truly happy with the results and excited to grow our partnership.
Clint Cowden
Assistant Superintendent/Vice President of Instruction & Student Services, Palo Verde College
We have been working with the N2N team for some time now and couldn’t be happier with their service and expertise. They’ve surpassed all our expectations.
Cem Erdem
CEO, AuguSoft Inc.
LightLeapAI ® virtually eliminated our fraudulent enrollment problem overnight, giving us real students, renewed faculty trust, and renewed hope for our college’s future. We’re truly grateful for their swift and effective partnership.
Dr. Jeannie Kim
President, Santiago Canyon College
The new fraud detection product doubled our ability to stop fraud, saves our teams time, and makes sure aid reaches real students. Integration was easy—even skeptics are now convinced. The product truly speaks for itself.
Dr. Jory Hadsell
Vice Chancellor of Technology, Foothill-De Anza Colleges
LightLeapAI ® automated our student services process, saving staff time, closing equity gaps, and ensuring students get quick access to the support they need. Our managers now respond instantly, making a real difference for our students.
Dr. Seher Awan
President, Mission College
I know how hard N2N team has worked to get us into position for the MyPath go-live. Please know we understand and respect the sacrifice of the N2N team to move BU forward.
Lee Johnston
Associate Vice Chancellor of Information Technology Services, Brandman University
Banner (R) Registration is a complex process and our system used to crash during critical registration periods. N2N's team assisted my campus with Performance testing the System and helped us create a reliable infrastructure for years to come.
Mia He
Director of Enterprise Systems, Wentworth Institute of Technology
The launch of the Brandman CBE Program is formally underway. I want to express appreciation to you and the N2N team for your leadership, commitment and good efforts to get us this far on our journey. Congratulations and many thanks, Kiran and team.
Phil Doolittle
Chief Financial Officer, Brandman University
Take Control. Create Opportunities. Reimagine Tomorrow...Today!
