RDS Supabase AI Application: A Managed BaaS Platform

What Is Supabase?

Supabase is an open-source alternative to Firebase (108k+ GitHub Stars, Apache-2.0 license). Built on PostgreSQL, it is a full-stack Backend-as-a-Service (BaaS) platform that provides complete backend capabilities, including PostgreSQL databases, user authentication, auto-generated RESTful and GraphQL APIs, real-time data subscriptions, object storage, and vector search. It enables developers to quickly build backend logic for web and AI applications without spending weeks setting up complex backends, databases, and servers. Unlike Firebase, which uses a NoSQL document database, Supabase leverages the power of the enterprise-grade relational database PostgreSQL. It combines the advanced features of relational databases with the flexibility of an open-source ecosystem, providing an open-source solution to avoid commercial BaaS vendor lock-in.

RDS Supabase applications are currently free during beta testing. After they become commercially available, this feature will be billed.

Supabase Capabilities

Supabase integrates various mature, best-in-class open-source tools into a unified development platform, providing core capabilities out of the box, such as data APIs, user authentication, object storage, and real-time communication. Developers can rapidly build high-performance, scalable applications without building backend infrastructure from scratch. Core capabilities include:

• PostgreSQL database
  • Provides an independent, fully managed PostgreSQL database with full SQL capabilities, rather than a simplified edition.
  • Supports JSONB, full-text search, views, stored procedures, triggers, and visual SQL tools.

• Automatically generated data APIs
  • When you create database tables, a Supabase application automatically generates secure, ready-to-use RESTful and GraphQL APIs based on the database schema.
  • The automatically generated RESTful APIs support database CRUD operations, making it easy to create, read, update, and delete data without writing backend code.
• User authentication (Authentication)
  • Provides out-of-the-box sign-up, sign-in, and session management, supports email-based sign-in and authentication, and uses JSON Web Tokens (JWTs) to implement fine-grained access control.
  • By tightly integrating user roles with PostgreSQL row-level security (RLS) policies, the system natively isolates data across users, ensuring that each user can access only their own data.
• Real-time subscription (Realtime)
  • Uses WebSockets to listen for database changes, making it easy to build features that require real-time updates, such as collaborative applications, chat, and dashboards.
  • Supports INSERT, UPDATE, and DELETE events, as well as Presence (online status) and Broadcast (message broadcasting).
• Storage
  • Integrates with Huawei Cloud Object Storage Service (OBS) to store and serve large files, such as videos and images.
  • Provides file-level access control based on RLS and supports resumable transfer of large files.
  • Provides image transformation and resizing with a built-in CDN.
• Vector and AI (Vecs/pgvector)
  • Includes the built-in pgvector extension, allowing vector embeddings to be stored and queried directly in PostgreSQL.
  • Eliminates the need for an additional vector database, simplifying AI application architecture.

RDS Supabase Architecture

RDS Supabase uses a layered, decoupled architecture for higher scalability and flexibility. Figure 1 shows the overall architecture.

Figure 1 Architecture

• Application Access Layer (ACCESS)

  The unified user entry point that provides multiple ways to access and manage the platform. This layer shields backend complexity to achieve single integration for multi-terminal use.

  • ManageAPI automates project management and orchestration.
  • OpenAPI exposes RESTful APIs for programs to call.
  • SDK provides multi-language development kits.
  • WebConsole serves as the visualized management console.
  • MCP supports AI Agent access.
• RDS Service Layer (SERVICE)

  The core service layer that is built on Supabase, responsible for routing, authentication, data access, and O&M capabilities.

  • Kong handles traffic control and routing.
  • Auth provides security authentication and user management.
  • PostgREST automatically converts API requests into SQL statements.
  • Realtime pushes database table changes.
  • Storage and Imgproxy handle large files and images.
  • Studio provides visualized O&M.
  • Supavisor handles high concurrency through connection pooling.
  • Logflare supports structured log analysis.
  • Newly added components include Meta (metadata management), Vector (vector search), and DBS Agent (management agent service). They extend the platform with metadata governance, AI vector retrieval, and control-plane interaction capabilities.
• RDS Business Data Layer (DATA)

  The business data storage and processing layer that is built on PostgreSQL. It uses a primary/standby architecture (primary node for read and write operations, standby node for read-only operations), offering high availability, reliability, and scalability. Relying on RLS policies, Auth schema, pgvector vector storage, and pggraph graph queries, it ensures data security and multi-modal query execution.

• Cloud Storage Layer (STORAGE)

  The underlying storage powered by Huawei Cloud. Huawei Cloud OBS provides S3-compatible object storage to host large files securely, while cloud disks provide persistence for workload data backups. The service layer accesses object storage through the S3 protocol and performs backup and recovery operations using backup protocols. The data layer communicates with the cloud storage layer through backup-and-recovery data flows, enabling secure data archiving and rapid recovery.

RDS Supabase Application Scenarios

• Rapid MVP Validation and Agile Delivery

  RDS Supabase enables products to move from prototype to production with minimal effort. After data modeling, backend capabilities are automatically generated from table schema definitions, eliminating the need to build services from scratch. Authentication, real-time messaging, and other common modules are available out of the box, allowing developers to focus on core business logic and significantly shorten delivery cycles.

• Backendless Web and Mobile Application Development

  Traditionally, building a complete application requires both frontend and backend systems. RDS Supabase adopts a "Database-as-an-API" model, exposing the data layer directly as callable APIs, significantly reducing backend coding and server O&M operations. Frontend engineers can deliver production-ready web and mobile applications without relying on server-side technology stacks (like Node.js) or worrying about deployment and scaling. This model is ideal for small teams and full-stack frontend engineers.

• One-Stop Foundation for AI Applications
  As AI becomes a standard product feature, integrating underlying data and model services becomes a major challenge. RDS Supabase is deeply integrated with PostgreSQL to provide end-to-end development support for AI applications.

  • Vectorized multi-modal storage: RDS Supabase relies on the pgvector capabilities of RDS for PostgreSQL to store vector data and perform similarity searches, providing the foundation for semantic matching. RAG applications do not require an additional vector database.
  • Unified APIs: RDS Supabase acts as the API layer between applications and data, shielding underlying complexity.
  • Scenario-driven capabilities: Common AI requirements, such as RAG retrieval and streaming WebSocket interactions, can be implemented using RDS for PostgreSQL and Supabase ecosystem capabilities. These capabilities support scenarios like intelligent customer service, smart Q&A, and semantic search, helping AI businesses build MVPs, validate market fit, and acquire early users with significantly lower effort.
  • AI knowledge base: By storing business data and AI vector data in the same database, RDS Supabase eliminates the need to maintain multiple heterogeneous systems. This avoids cross-system data synchronization challenges and significantly reduces the development and O&M costs of AI applications.
• SaaS Multi-Tenancy and Fine-Grained Access Control

  For SaaS providers and enterprises undergoing digital transformation, data and permission isolation is a foundational requirement. RDS Supabase provides an Auth mechanism supporting multiple authentication methods and combines it with RLS policies to deliver fine-grained control over data access. This ensures strict tenant-level isolation in multi-tenant scenarios while supporting role- and dimension-based privilege division within organizations, enabling security governance to extend across the entire lifecycle of the application.

• Enterprise-Built Internal BaaS Platform
  • It standardizes the enterprise's backend architecture, eliminates redundant development, and ensures full data sovereignty.
  • Its visualized management console simplifies platform operations, enables isolation across multiple projects, and mitigates data silo challenges within the enterprise.
• Real-Time Collaborative Applications

  Real-time performance is core to user experience of collaborative products. Realtime of RDS Supabase provides capabilities such as data change listening (Postgres Change), message broadcasting (Broadcast), and online status tracking (Presence), covering key collaboration scenarios. Developers can efficiently build applications based on these capabilities and extend them to enterprise-grade application scenarios, including online multi-use document editing, real-time updates of project management dashboards, and message and status synchronization in instant messaging systems.

Advantages of the RDS Supabase Development Model

Traditional application architectures are typically multi-layered and tightly coupled, with frontend services, backend services, and databases closely bound together. Each component needs to be independently developed, deployed, and maintained. Supabase breaks this bottleneck by offering integrated services.