April 2026 Summaries

Firebase Data Connect has been renamed Firebase SQL Connect to emphasize its expanded PostgreSQL capabilities and its goal of combining Firebase’s developer experience with relational databases. New features include real-time query updates through an @refresh directive, offline caching at query and entity levels, and Native SQL support for advanced PostgreSQL use cases such as geospatial queries, window functions, and complex transactions, with deployment-time validation and server-side parameterization. The platform also introduces experimental Custom Resolvers using Cloud Functions for Firebase, enabling applications to integrate external APIs, Cloud Storage, BigQuery, and other data sources while retaining unified authorization and type-safe client SDKs. Pricing updates include 90-day no-cost PostgreSQL trials on the Spark plan without a credit card, expanded free trials for Blaze users, and operation pricing starting at $0.90 per million operations with 250,000 free monthly operations on Blaze. SQL Connect additionally builds on previously released full-text search, type-safe Admin SDKs, and event triggers for server-side workflows.

Firestore Enterprise pipeline operations have reached general availability, expanding the database’s query capabilities with subqueries for cross-collection joins and aggregations, plus data manipulation language stages for targeted updates and deletes. New full-text search and geospatial querying features remain in public preview, enabling relevance-ranked document search with text indexes, synonym and spelling handling, and location-based results ordered by distance. Pipeline stages can be combined to filter, sort, aggregate, and modify data in a single Firestore workflow, such as finding compatible agent skills, calculating restaurant review summaries, or backfilling missing document fields. Updates and deletes are currently supported through DML, while inserts and transactions are planned, with the release positioned as bringing Firestore closer to the query flexibility associated with relational databases while retaining its NoSQL and serverless design.

At Cloud Next 2026, Firebase announced updates intended to make Google Cloud services more accessible to mobile and web developers while improving application scalability, security, and AI integration. Firebase Data Connect is becoming Firebase SQL Connect, adding native SQL, real-time synchronization, offline caching, and Cloud SQL for PostgreSQL support through a no-cost trial. Firestore Enterprise is expanding its query capabilities with full-text search, geospatial queries, subquery-based joins, and data manipulation features, while Firebase’s integration with Google Cloud Application Design Center provides centralized deployment, governance, and AI-assisted infrastructure management. Firebase AI Logic is adding tools such as automatic function calling, JSON schema mapping, secure prompt templates, replay protection, hybrid on-device and cloud inference, and content caching to simplify and control generative AI features. Other announcements include experimental Dart support for Cloud Functions, general availability of SMS-free Firebase Phone Number Verification in May 2026, and Firestore and Firebase Authentication integration with Google AI Studio for building full-stack applications through natural-language prompts.

Firebase AI Logic announced several Cloud Next ’26 updates aimed at improving security, prompt management, cost efficiency, and developer workflow while retaining its client-first approach to Gemini integration for mobile and web apps. Server prompt templates now support multi-turn chat and function calling, allowing system instructions, model configurations, tool schemas, and safety settings to be centrally managed in Firebase without requiring app updates, while clients manage conversation flow and execute local functions. Firebase App Check will introduce single-use token replay protection in May 2026 to reduce abuse of AI endpoints, and SDKs now offer type-safe automatic function calling that derives schemas and handles data serialization from native code structures. Explicit context caching lets developers reuse large, previously uploaded content through cache IDs to reduce token costs and latency, particularly when combined with server-managed templates. Firebase is also introducing experimental hybrid on-device inference for Android, which uses Gemini Nano where supported for local, private, no-cost inference and otherwise falls back automatically to cloud models.