Proof

Powered by MetaObjects.

A selection of production systems built on MetaObjects — my own consulting engagements, the Draagon AI products I build on it, and the earlier work that came out of two decades of metadata-driven development.

Open-source reference implementation

Wizards of Odd

TypeScript / React / MetaObjects

A live AI wizard-council app built to be read: one metadata spine drives the database schema, API types, LLM prompts, and SSE streaming protocol — rename one field and all of them fail in CI. Apache-2.0, clone-and-read.

AI agent platform

Draagon AI

Python / TypeScript / React / MetaObjects

From 17 runtime schema patches to zero: one ~7,000-line YAML spine generates more than 20,000 lines of drift-checked Python and TypeScript — 44 entities and more than 100 LLM prompts declared once, on the published MetaObjects packages.

Voice assistant · the small-adoption case

Roxy

Python / Postgres / MetaObjects

Not every adoption is a rewrite. An existing Python voice assistant took exactly two of the six pillars — prompts rendered from typed payloads at runtime, and Postgres schema owned by the CLI — in about seventy lines of metadata, and left the rest of its codebase alone.

AI Game Master platform

Party Lore

Java / Kotlin / TypeScript / MetaObjects

288 hand-written schema migrations collapsed to one generated baseline, ~1,900 lines of hand-built LLM prompts became typed contracts, and the model made deleting dead schema mechanical — 29 tables and 49 columns removed in nine days, each proven safe by a recompile.

Smart Monitor · open-source system monitor

smon

TypeScript / Bun / MetaObjects

A 377-line bash system monitor rewritten into a typed, metadata-driven TypeScript project. Probes, verdicts, alert payloads, email templates, and notify adapters are declared as MetaObjects metadata; the registry, type unions, config checks, and docs are generated from it. Honest before/after — including what wasn't clean.

High-concurrency poker platform, now an AI benchmark

Online Poker Platform → FeltBots

Java 8 / TypeScript (Bun) / Python / MetaObjects

Built rapidly on MetaObjects in the early 2000s; the engine survived two decades and now runs FeltBots, where language models play for measurement rather than money. Its 155-message wire protocol — undocumented for most of that time — is now declared once as metadata, and the TypeScript bridge, the Python bots and the protocol docs are generated from that declaration.

First TypeScript MetaObjects reference

Michael Frank Training

TypeScript / React 19 / Fastify / Drizzle / MetaObjects

The first reference project for the new TypeScript MetaObjects. A modern training and subscription platform at officialmichaelfrank.com — React 19 + Fastify monorepo with subscription management, all driven by MetaObjects schema.

Pharmaceutical regulatory platform

Liquent Insight (PAREXEL)

Java / MetaObjects

Customizations per pharmaceutical customer with zero code changes — accomplished by overlaying customer-specific metadata on the base model. Model extensions, UI behaviors, grid layouts, access controls, and validations all reconfigured per tenant by metadata alone.

Salesforce-integrated secure messaging

Sendside Networks

Java / Salesforce / MetaObjects

Secure messaging and digital contract platform, integrated with Salesforce. Built rapidly on MetaObjects — multiple teams shipped in parallel against the same metadata patterns. Dynamic forms, e-signatures, and confidential data exchange, RESTful integration suite throughout.

Content system at Cengage

Logical Content Store

Java / MetaObjects

Content deployment pipeline treating content like source code: advanced merging, real-time preview, regenerated derivative artifacts. The first implementation leveraged MetaObjects.

Anonymized — fintech client

Credit Card Processing Gateway

Java / MetaObjects

Production payment-processing gateway built rapidly on MetaObjects — same rapid-development pattern as the poker platform and Sendside. Metadata-driven transaction routing, validation, and audit.

By the numbers

One metadata approach — from a two-day open-source tool to a two-decade production engine.

4
languages generated from one model, across these projects
Java · Kotlin · TypeScript · Python (the standard's conformance suite targets a fifth, C#)
20+ yrs
the approach in production — from the 2000s poker engine to today
2
open-source references you can clone and read now — smon & Wizards of Odd
0
silent schema/type drift — every reference gates the model against the code in CI
9
projects ran the release candidate with their drift gates enforced before 1.0 was cut — the round found five issues, fixed before the stable release rather than after it

Every figure is verified per project — see each case study for its sources. The nine projects above are all maintainer-owned. That is stronger evidence than a test suite and weaker than independent adoption.

Considering MetaObjects for your next platform?

NIX Solutions can review your architecture, run a pilot or help with migration.

Contract with NIX ↗