What people build with dbo.io
dbo.io is a general-purpose framework. But some patterns come up again and again — here are the ones it handles especially well.
Give your agents infrastructure
they can actually work with
AI agents need to read data, write data, and understand what data is available — without an engineer manually wiring up each capability. Most API layers are built for human clients: static endpoints, opaque data models, security enforced in application code that agents cannot inspect.
How dbo.io fits
Discovery built in
The /api/app/object/{appName} endpoint returns the complete application definition as structured JSON — every entity, every named query, every security rule. An agent can locate the right endpoint by name, description, or UID without human intervention.
Schema introspection
/api/output/meta/entity/{uid} returns full column metadata: types, constraints, relationships, encryption flags, display names. Agents understand the data model without reading source code.
Dry-run writes
Every write through /api/input/submit is a two-step operation: validate first (no _confirm=true), commit second. Agents can propose a write, review the validation result, and only commit when the operation is confirmed safe.
Cell-level security
Entity, row, and column access controls are enforced at query time — before SQL executes. An agent with access to invoice records cannot see payroll records, even if they share a database.
Agent identities
Agents authenticate as distinct identities and receive security grants the same way human users do. Configure what an agent can access through the same security data records that govern human access.
Audit trail
Every operation — read or write — is logged in the revision system. Agent actions are reversible. Compliance requirements apply equally to humans and agents.
Function calling
The chatbot integration routes AI tool calls through the platform by naming convention: a function named output_orders resolves to /api/o/orders. Function arguments flow through the standard token pipeline. No custom routing code required.
Who this is for: Teams deploying autonomous agents for tasks that involve reading from or writing to structured business data — data retrieval, report generation, record creation, workflow automation.
Ship internal tools that your
compliance team will also love
Internal tools accumulate technical debt faster than any other category of software. They start as quick prototypes, grow into critical dependencies, and end up maintained by whoever still understands the codebase. Security is bolted on. Audit trails are afterthoughts.
How dbo.io fits
No middle tier to drift
When a business requirement changes, you update a metadata record — a column definition, a security rule, a query filter. The API reflects the change immediately. No controller to update, no ORM to regenerate.
Cell-level security from day one
Role-based, row-level, and column-level access control are configured as data records and enforced automatically. Finance users see financial records. Configure it without writing permission logic.
Revision history on everything
Every change to every configurable asset writes a revision record with the user identity, timestamp, and full snapshot. Roll back to any prior state. Satisfy audit requirements without building an audit system.
App portability
Export an internal tool as a .model archive. Promote from dev to staging to production as a single import. No environment drift, no manual sync.
Who this is for: Organizations building internal tools, admin dashboards, operations management systems, or employee-facing applications where security, auditability, and long-term maintainability matter.
Multi-tenant by design, not by retrofit
Adding multi-tenancy to an existing application is expensive and error-prone. Tenant isolation that lives in application code rather than the data layer is one bug away from a data breach.
How dbo.io fits
Infrastructure isolation
The Account → Instance model gives each tenant a physically isolated database. Data separation is guaranteed at the infrastructure layer, not enforced through application logic.
Per-tenant config as data
Security rules, query definitions, content templates, and user groups are all data records. Configure them differently per instance without branching code. One codebase, unlimited tenants.
App distribution model
Package a SaaS application as a .model archive. Provision a new tenant by importing the archive into a fresh instance. Version releases are tracked with per-asset change records.
Who this is for: Teams building SaaS platforms that require strong tenant isolation, per-tenant configuration, and a consistent API surface across tenants.
Add a modern API layer to an
existing system without touching it
Legacy databases often hold critical business data but were never designed to be accessed via modern APIs. Rebuilding from scratch is too expensive and too risky. Wrapping the existing database in a new application layer is the only practical option — but it requires building and maintaining a full middle tier.
How dbo.io fits
Point dbo at your existing database
The Data Source Sync API imports the schema — all tables, columns, types, and relationships — into dbo's entity metadata. No schema redesign, no data migration.
Full dbo API surface immediately
Query endpoints, write endpoints, schema introspection, security enforcement. No code to write. dbo reads from and writes to the external database using standard SQL without modifying the schema.
Add what's missing
REST API for mobile clients, cell-level access control, revision history, content templates for web interfaces, messaging for notifications, scheduled automation for background processing.
Example
An enterprise organization uses their existing operational database as a data source, syncs the schema into dbo, and builds a mobile application and customer portal on top of it — without modifying their transaction processing system. The legacy system continues to process transactions; dbo adds the API, the mobile app, and the reporting layer.
Who this is for: Organizations with existing databases that need to extend them with modern API access, mobile/web interfaces, or AI integrations — without rebuilding the underlying system.
One API. iOS, macOS, web, and AI agents.
All at once.
Shipping a cross-platform application typically means maintaining separate API layers, separate authentication systems, and separate permission models for each client type. As the number of clients grows, the surface area for inconsistency grows with it.
How dbo.io fits
One API, every client
The dbo REST API is the only backend your applications need. iOS, Android, web, macOS, CLI tools, and AI agents all authenticate to the same endpoints and operate within the same security model.
No API versioning overhead
Because the API surface is generated from schema metadata, adding a field to the data model doesn't require a new API version. The change propagates automatically. Existing clients that don't request the new field are unaffected.
Authentication for every context
Username/password, email/password, phone/passkey, HTTP header SSO, SAML 2.0 — every authentication pattern is built in. Mobile apps, web sessions, and server-to-server integrations each use the method appropriate for their context.
MCP-ready
dbo-generated APIs are natural targets for Model Context Protocol servers. The Queue application ships with an MCP server that exposes its API to AI assistants. The combination of dbo and MCP gives AI agents structured, secure access to application data.
The Queue application is a social link-sharing tool built entirely on dbo.io. It has a dedicated iOS app, a macOS app, a web app, an iOS share extension, and an MCP server. All of them run against the same dbo-generated API. No separate API layer was written for any of these clients.Queue — built by dbo.io as a reference implementation
Who this is for: Teams shipping mobile-first or cross-platform applications who want a single, consistent API backend that scales to any client type — including AI agents.
Ready to stop writing the middle tier?
dbo.io is invite-only. We work with a small number of early partners to make sure the integration is set up right. Tell us about what you're building.