Two ways in: a coding agent runs the Norbix CLI, a chat app calls the MCP server every Norbix installation already is. Both act with exactly the roles you pick — and build, configure, and run a production backend through conversation.
A coding agent has a shell, so it runs the Norbix CLI. A chat app has none, so it calls the MCP server every Norbix installation already is. Both act as an AI service user with exactly the roles you pick.
norbix login → norbix ai init → norbix … --json --dry-run first, --yes only for destructive changes. Same commands in CI.https://hub.norbix.ai/v3/account/mcp Copy-paste setup for every tool on Integrations.
See how ChatGPT helped build MotoSkelbimai and Claude Code helped rehearse Happy Tummy’s backend migration.
Marketplace · Web app Working public portal
See how a motorcycle marketplace uses Norbix for searchable data, seller accounts, moderation and its email engine—with ChatGPT helping build the custom website.
ChatGPT + Norbix CLI
See the modules behind it
Parenting · Mobile app Migration tested in TEST
How a small team is bringing an existing parenting app to Norbix—preserving accounts, sharing content across three languages and reusing backend tools instead of building them.
Claude Code + Norbix CLI
See the modules behind itPick your tool. See how it builds on Norbix.
Tell Claude what you need. Get a production backend.
Claude Code signs in once with the Norbix CLI (norbix login, norbix ai init), then scaffolds schemas, configures auth, wires events, and deploys modules — through conversation, one JSON answer per command.
View integrationCodex reads AGENTS.md. Norbix writes it.
norbix ai init --target agents puts the Norbix block into AGENTS.md. Codex then runs norbix commands against your backend — or connects to the hosted MCP server with codex mcp login.
View integrationBuild your backend without leaving your editor.
The Norbix CLI writes a rules file for Cursor's agent; the hosted MCP server works too. Ask it to set up your database, configure payments, or wire a signup flow — it scaffolds production-ready code against real modules.
View integrationFrom prompt to production in one session.
Windsurf's Cascade reaches Norbix through the hosted MCP server, or runs the Norbix CLI from its terminal. Auth, database, email — configured and deployed through conversational coding.
View integrationYour backend as a connector, everywhere Claude runs.
claude.ai, Claude Desktop and Cowork share one custom connector pointed at the hosted MCP server. Sign in with OAuth, pick the roles, and talk to your backend in any chat.
View integrationChatGPT asks, your backend answers.
Add a connector with the MCP address and OAuth. ChatGPT opens Cloud's consent screen, you pick the roles — for example Viewer on one project — and press Allow.
View integrationOpen-source agents. Production-grade backends.
OpenClaw agents use Norbix as their action layer — scaffolding backend infrastructure, managing data, and triggering workflows without proprietary lock-in.
View integrationAutomate anything. Norbix handles the backend.
n8n calls Norbix from an HTTP Request node with a service-user key, and Norbix webhooks start n8n workflows on signups, payments, or any event. Automations that read and write to your production backend.
View integrationStateful AI agents with a real backend.
LangGraph agents treat Norbix as a first-class tool — querying databases, creating users, sending emails, and processing payments in multi-step agent workflows.
View integrationChain LLM calls to production backend actions.
LangChain tools connect directly to Norbix modules. Your LLM pipeline can create records, trigger emails, check permissions, and manage files — against a real, secure backend.
View integrationGive your AI crew a real backend to work with.
CrewAI agent teams use Norbix as their shared action layer. One agent handles data, another manages payments, a third sends notifications — all against the same backend.
View integrationRapid prototyping meets production infrastructure.
Fizzy's fast iteration paired with Norbix means your prototypes aren't throwaway — they're built on production modules from the start.
View integrationWhat makes Norbix different from letting your AI generate a backend from scratch.
AI-generated backends are fragile — no auth model, no audit trail, no event system. Norbix enforces security and structure. What your AI produces is production-ready from the first run.
The same Norbix your AI builds against at development time is the Norbix everything calls in production. No second platform. No rewrite. No handoff.
All 9 modules share one event bus. A user signs up → welcome email fires → Stripe customer is created → onboarding workflow triggers. Your AI wires it. Norbix runs it. Zero glue code.
Role-based access, audit trails, encryption, and data ownership are built in — not bolted on. Even if your AI doesn't think about security, Norbix does.
From zero to production in three steps
Coding agent: npm i -g @norbix.ai/cli, norbix login, norbix ai init. Chat app: add https://hub.norbix.ai/v3/account/mcp as a connector and sign in with OAuth. Pick the roles, press Allow.
Tell your AI to set up auth, create a database schema, wire a payment flow, configure email templates — in plain English.
Deploy on your cloud or ours. What your AI built is production-ready — secure, event-driven, and connected to all 9 modules.
Create a free account, connect your favourite AI tool, and ship a production backend today.
Join the waitlist