// the find
steedos/steedos-platform
The AI-Native Infrastructure for Enterprise Apps. Powered by ObjectStack (ObjectQL, ObjectOS, Object UI). Turn Prompts into Enterprise Software.
Steedos is a Salesforce-style low-code platform for enterprise apps, built on a metadata-driven architecture where YAML schemas define objects, UI, and workflows. It targets teams who want to build CRM/ERP-class apps without writing everything from scratch. The project is mid-rebrand from 'Steedos' to 'ObjectStack', which means documentation, code, and naming are currently inconsistent.
The metadata-as-code approach (*.object.yml, *.page.yml) is genuinely useful — it makes apps version-controllable and AI-generatable in a way drag-and-drop tools are not. Field-level RBAC built into the schema layer is a real feature that most low-code tools bolt on as an afterthought. The NestJS + Moleculer microservices backend is a reasonable architectural choice for multi-tenant enterprise workloads. Docker quickstart actually works: one command to a running instance, which is rare for platforms this complex.
The repo root is littered with AI-generated analysis files (CODE_QUALITY_ANALYSIS.md, EXPLORATION_SUMMARY.txt, QUICK_ACTION_ITEMS.md) that have no business being committed — signals the codebase is being navigated by AI tools, not maintained by humans with clear ownership. The mid-migration state is a real adoption risk: the README explicitly says future development is on ObjectStack but you still deploy Steedos, and the two naming systems coexist without a clear compatibility story. MongoDB is the default backing store despite the README claiming 'database agnostic', and switching to Postgres mid-project is non-trivial. The AI features are thin wrappers around LLM calls for schema generation — there is no validation loop, so generated YAML that doesn't match the schema spec just silently fails at runtime.