DEV Community

Mihaela Gheorghe-Roman
Mihaela Gheorghe-Roman

Posted on

A Closer Look at Oracle Backend with Firebase APIs

#oracle #webdev #database #firebase

Oracle recently launched Oracle Backend with Firebase APIs - a toolkit for building mobile and web apps on Oracle AI Database, with open source SDKs for auth, document data, storage, security rules, vector search, and App Trust.

The "Firebase" part refers to the developer experience, not the vendor. Oracle is explicit about this in the SDK repo: the naming exists to make the API feel familiar to Firebase developers, nothing more. The shape of it backs that up - a Console for configuring services, a CLI, client SDKs for iOS, Android, JavaScript and Flutter. What's different is everything underneath. Instead of Google's infrastructure, it all runs on Oracle AI Database, with ORDS 26.1.1 exposing it as REST endpoints the SDKs call.

Feature-wise you get what you'd expect from a BaaS: authentication (email/password, social, enterprise IdPs), a document-style database with collections and subcollections, file storage, and security rules enforced at the database level.

Vector search story

The most relevant piece here is the vector search story. Instead of standing up a separate vector database - which is what most teams end up doing the moment they need embeddings or semantic search - the vectors sit in the same Oracle AI Database as everything else. Same tables, same security rules, one query that joins a findNearest result straight into relational data.

Anyone who has dealt with keeping a vector store in sync with the source of truth will recognize why that's not a minor detail. It removes a whole category of operational work: no second system to provision, secure, back up, and keep consistent with the primary one. For teams already on Oracle that need semantic search over support tickets or a product catalog, this closes that gap without adding a new moving part to the stack.

One honest caveat: this isn't a managed, Oracle-hosted service. You still stand up and run Oracle AI Database and ORDS yourself - OCI, another cloud, Cloud@Customer, on-prem. Fusabase — the name used for the CLI, the npm package, and the GitHub repos, and it shows up on the official landing page too — is the layer on top. Depending on how much you value staying inside infrastructure you already control, that's either the whole point or the catch.

If you want to try it

Oracle has three LiveLabs workshops - iOS, Android, Web - that all build the same demo app, RecipeShare, covering SDK setup, the Console, auth, security rules, database and storage. If you don't already have a mobile dev setup ready to go, Web is the one you can easily use - you can work through it from any laptop, no Xcode or emulator involved.

Further reading

If this is up your alley, these are worth a real look:

Top comments (0)