Platform release

API v2: the platform, rebuilt

The biggest technical change since launch. We migrated the entire backend from the original PHP and CodeIgniter API to a new one on NestJS and TypeScript, ran it beside the old platform through a thirteen-phase rebuild, and cut production over on September 10. Same platform, new engine.

July 27 to September 10, 2026

Our v1 API got us here. It let us build the product, validate the workflows, serve customers, and merge two products into one. But as Octelligence grew to span corporate records, cap tables, equity instruments, stakeholder access, and portfolio management across our web and mobile apps, we wanted the foundation underneath all of it to match where we are taking the platform next. Over the summer we rebuilt that foundation, and on September 10 we cut the production platform over to it. This was not a prototype or a side project. We migrated the platform your team uses every day.

What changed

  • The backend moved from PHP and CodeIgniter to a new API on NestJS and TypeScript
  • A redesigned data model across every domain: identity, registry, securities, instruments, governance, documents, compliance, and audit
  • Authentication rebuilt so the new API issues its own tokens, with 2FA enforced at issuance
  • Document rendering ported to TypeScript, proven first on the share certificate
  • Customer data migrated across, then a sequenced, gated cutover on September 10
  • Nothing changes for customers: same records, same URLs, no action needed

Why we rebuilt it

v1 was a PHP and CodeIgniter application that carried the product from its first corporate record to a merged platform serving records and equity together. That is a long way for a codebase to travel, and by mid-2026 the constraints of the original architecture were showing up as friction: features that touched several domains were slower to build, the data model had accreted workarounds, and the web and mobile apps did not always see the platform the same way. A strongly typed API with a clean domain model fixes all three. It is faster to build on, easier to keep consistent across applications, and it gives us room to scale without carrying forward the compromises we started with.

How we did it without a big bang

Rewriting a production backend is where products go to break. We avoided the all-at-once cutover that causes that. Starting July 27 with the NestJS foundation for the new API, we worked through a rebuild plan of thirteen sequenced phases, each one redesigning and reimplementing a slice of the platform against measured data from the live system rather than against assumptions:

  • Identity and tenancy first: the account, corporation, and party model, then access and token issuance, so everything above it stood on a settled base.
  • The records and equity core: the registry and directory, securities and ownership, capital instruments, governance, the minute book, and compliance, each rebuilt as its own module.
  • The cross-cutting systems: billing as the single owner of subscription state, document rendering in TypeScript, audit, and the stakeholder portal views.

Throughout, the new platform ran beside v1 rather than in place of it. We migrated customer data across with extract-and-load steps that normalised it to the new model, then rehearsed the switch against a written cutover runbook: sequenced, gated at each step, and owner-tagged, so the actual cutover on September 10 was a controlled operation rather than a leap.

What customers see

As little as possible, which is the point. Records, cap tables, workflows, certificates, and both the Ledger and Equity Vault apps keep working exactly as before. Existing URLs continue to resolve. There is nothing to migrate on your end and nothing to reconfigure. The day-to-day experience is the same; what changed is underneath it.

The first mile on the new engine

The proof that the foundation is ready to build on is what we shipped the day after cutover. On September 11, in-app support went live in Ledger: file a ticket, track it, and hold a threaded conversation with our team without leaving the app. It is the first customer feature built natively on API v2, and it came together quickly precisely because the new architecture is faster to build on. It is covered alongside the summer's other work in the Summer 2026 release.

What's next

With v2 carrying production, feature work moves back to the front of the platform. The clean domain model and strong typing make the next wave of features faster to ship and easier to keep consistent across the web and mobile apps, and the new infrastructure gives us headroom to scale as more corporations and firms come on. There is a lot more coming. Same Octelligence, new engine.

By the numbers

  • 13 sequenced rebuild phases from foundation to cutover
  • 2 stacks retired and adopted: PHP and CodeIgniter out, NestJS and TypeScript in
  • 1 gated production cutover, on September 10
  • 0 things for customers to do

For the line-by-line view of the platform changes that led up to this, including the single API behind Ledger and Equity Vault and the split into two front doors, see the changelog.

Same Octelligence. New engine.

Corporate records and equity in one platform, now on a foundation built to scale. If that's the gap you're trying to fill, we'd like to help.

Start free trial See what shipped this summer