Skip to content
An AI-native data and rules platform

Shared rules.
Individual freedom.

Axiomony gives every employee control over their own user experience. Use your personal AI agent to design and refine forms, dashboards and workflows that match your exact tasks, needs and preferences, while the database keeps shared data and rules consistent.

Early development preview · Database-enforced rules · Model Context Protocol

Shared coreData · Rules · Access
My reports
My forms
My workflow
My dashboard
One foundation. Endless ways to work.
Why Axiomony exists

Stop fitting your business into someone else's software.

Traditional software development divides the world into frontend, backend, and database. Schema-first and low-code products can build on a data model, but you still inherit the vendor's choices about how the software looks, behaves, and integrates.

Creating or changing software across those layers takes technical and operational expertise. Developers make compromises for different departments, and overlooked requirements or new opportunities can become expensive change requests.

Axiomony is designed to remove those barriers. People who understand the organization can use an AI assistant such as ChatGPT, Gemini, or Claude to describe the data and binding rules: who can view or change what, when, why, and under which conditions. Authorized administrators review and approve changes that Axiomony applies to the database.

01

Define the shared truth

Model the organization's data, relationships, permissions, and business rules with AI assistance—not with a fixed set of application screens.

02

Enforce rules at the source

The database enforces permissions, integrity constraints, and business rules regardless of the interface. A generated tool cannot grant itself additional database rights.

03

Own your experience

Each person or department can create a tailored form, mobile workflow, dashboard, report, or authorized integration for exactly the work they do.

Purchasing might design a phone-friendly invoice scanner. Finance might connect an approved banking API using its own authorized credentials. Teams can adapt without waiting for a vendor's next release. New colleagues can start prototyping useful improvements in minutes.

The promise is operational autonomy within shared rules. Credentials and integrations still need secure handling, and this project is an early-stage prototype—not a production-ready deployment.

A practical example

One invoice policy.
Many personal workflows.

Suppose your organization allows invoices below €500 without a separate approval. That rule is enforced by the database, not hidden in a particular invoice screen.

  • An accountant creates a streamlined invoice-entry form.
  • A manager asks for a dashboard of invoices needing approval.
  • An auditor builds a report with the data their role is permitted to see.

Everyone uses the same data and rules—even when their interfaces look completely different.

Concept · approval policy
// Defined once, enforced by the database
IF invoice.total < 500 EUR
  approval_required = false
ELSE
  approval_required = true

// These are user-created experiences
Invoice entry    • Finance
Approval queue   • Manager
Monthly report   • Auditor
How it fits together

The database decides. Users design.

Two MCP capabilities serve different jobs. The final authorization decision always belongs to the database, not to a generated user interface.

AI assistantPlain-language request
→
Axiomony MCPData tools or admin planner
→
DatabasePermissions, constraints, rules
Know the boundaries

Built to explore—not yet built to expose.

Axiomony v0.5.1 is a development prototype. The design separates data authority from the interface, but public authentication, application sessions and production-safe AI migrations are not complete.

Local development only. Do not expose the MCP endpoint to the public internet. The current server authenticates using reusable database credentials over HTTP Basic; production OAuth/OIDC, rate limiting and complete integration tests remain future work.

Start building

One model. Your own way to work.

Start with a bundled database or connect an existing database. Then explore the tools available to your database account.

Installation guide →