firebase-ai-logic-basics

作成者: firebase

Firebase AI Logic(Gemini API)をウェブアプリケーションに統合するための公式スキル。セットアップ、マルチモーダル推論、構造化出力、セキュリティをカバーします。

npx skills add https://github.com/firebase/agent-skills --skill firebase-ai-logic-basics

Firebase AI Logic Basics

Overview

Firebase AI Logic is a product of Firebase that allows developers to add gen AI to their mobile and web apps using client-side SDKs. The app can directly call Gemini models via Firebase AI Logic without the developer needing to manage a dedicated backend. Firebase AI Logic was previously known as "Vertex AI for Firebase".

It supports the two Gemini API providers:

  • Gemini Developer API: It has a free tier ideal for prototyping, and pay-as-you-go for production
  • Agent Platform Gemini API (formerly branded Vertex AI): Ideal for scale with enterprise-grade production readiness, requires Blaze pricing plan

Use the Gemini Developer API as a default, and only Agent Platform Gemini API (formerly branded Vertex AI) if the application requires it.


Setup & initialization

Prerequisites

  • Identify the platform the user is interested in building on prior to starting: Android, iOS, Web, or Flutter.
  • If their platform is unsupported, Direct the user to Firebase Docs to learn how to set up AI Logic for their application (share this link with the user https://firebase.google.com/docs/ai-logic/get-started)
  • The Firebase CLI commands below run through npx on every platform, so they need Node.js 20+ and npm. Use the firebase-basics skill to check for them and to log in to the Firebase CLI. If Node.js is missing or older than version 20, ask the user to install it as that skill describes. Don't install system packages yourself, for example with sudo apt-get, because that changes the user's machine without their consent.

1. Provision Firebase AI Logic (ALL platforms)

Use the Firebase CLI to provision Firebase AI Logic for the Firebase Project:

  1. Make sure that you're in a firebase directory (with a firebase.json).

  2. Verify that you're running Firebase CLI commands against the intended Firebase Project. When you run this command, the active project is marked with "current":

    npx -y firebase-tools@latest projects:list
    
  3. Make sure there's at least one Firebase App for the target platform registered with the current Firebase Project:

    npx -y firebase-tools@latest apps:list
    
  4. Provision Firebase AI Logic for the Firebase Project:

    npx -y firebase-tools@latest init ailogic
    

    This command will enable the Gemini Developer API in the Firebase project.

[!WARNING] CRITICAL: Provisioning Firebase AI Logic is Required For all platforms (Android, iOS, Web, or Flutter), even if the app already uses Firebase, you MUST run npx -y firebase-tools@latest init ailogic to provision the service. The configuration files of GoogleService-Info.plist, google-services.json, and firebase_options.dart ONLY handle client configuration and do NOT enable the Firebase AI Logic service, leading to PERMISSION_DENIED errors.

If you can't run init ailogic yourself (for example because Node.js is missing or the Firebase CLI isn't logged in) do NOT skip it. The app's requests to Gemini models will fail until the Firebase AI Logic service is provisioned. Tell the user to run this command in their project directory, and include it in your final summary:

npx -y firebase-tools@latest init ailogic

2. Add the SDK and initialize the service in the app

Adding the SDK dependency and initializing the service are platform-specific. Before writing code, read the reference for the target platform listed under Initialization Code References.

More info in Firebase AI Logic Getting Started


Core Capabilities

[!WARNING] CRITICAL: Use current model names: Always check the Firebase AI Logic Models documentation for the currently supported model names. Do NOT use any Gemini 1.x or Gemini 2.x models (such as gemini-1.5-flash, gemini-1.5-pro, gemini-2.0-flash, gemini-2.0-pro, gemini-2.5-flash, gemini-2.5-flash-lite, or gemini-2.5-pro), as older models are deprecated or shut down. Always use a currently supported Gemini 3+ model; if URL access is blocked in your environment, proceed directly without searching local files.

Text-Only Generation

Multimodal (Text + Images/Audio/Video/PDF input)

Firebase AI Logic allows Gemini models to analyze image files directly from your app. This enables features like creating captions, answering questions about images, detecting objects, and categorizing images. Beyond images, Gemini can analyze other media types like audio, video, and PDFs by passing them as inline data with their MIME type. For files larger than 20 megabytes (which can cause HTTP 413 errors as inline data), store them in Cloud Storage for Firebase and pass their URLs to the Gemini Developer API.

Chat Session (Multi-turn)

Maintain history automatically using startChat.

Streaming Responses

To improve the user experience by showing partial results as they arrive (like a typing effect), use generateContentStream instead of generateContent for faster display of results.

Text-to-Speech (TTS) Generation

Generate spoken audio directly on client devices without a custom speech backend. Firebase AI Logic supports speech synthesis using dedicated Gemini TTS models:

  • Supported Models: gemini-3.1-flash-tts-preview
  • Capabilities:
    • Single-speaker voice persona selection (voiceName) and multi-speaker dialogues (up to 2 distinct speakers) via SpeechConfig / MultiSpeakerVoiceConfig
    • Direct audio responses by setting responseModalities to audio
    • Streaming speech responses with generateContentStream for low-latency playback
    • Audio directives ([Audio Profile: ...], [Scene: ...], [Director's Note: ...]) and emotional tags ([whispers], [laughs], [slowly])
    • Client-side decoding and playback of 24 kHz 16-bit linear PCM (audio/l16) or WAV container format (audio/x-wav)

See the Text-to-Speech (TTS) Generation Guide for full implementation details, streaming patterns, and copy-paste code snippets for iOS (Swift), Android (Kotlin), and Web.

Generate Images with Nano Banana

[!WARNING] Use current Image model names: Always check the Firebase AI Logic Models documentation for the currently supported image generation (Nano Banana) model names.

  • Requires an upgraded Blaze pay-as-you-go billing plan.

Search Grounding with the built in googleSearch tool


Supported Platforms and Frameworks

Supported Platforms and Frameworks include Kotlin and Java for Android, Swift for iOS, JavaScript for web apps, Dart for Flutter, and C Sharp for Unity.

Advanced Features

Structured Output (JSON)

Enforce a specific JSON schema for the response.

On-Device AI (Hybrid)

Hybrid inference runs prompts with an on-device model when one is available and falls back to a cloud-hosted model otherwise (or vice versa). Availability and setup steps differ by platform, more info in the hybrid and on-device documentation, which links to each platform's guide.


Security & Production

Firebase App Check

[!WARNING] Critical Safety Requirement: In order to use AI Logic safely, you MUST set up App Check on your app. This prevents unauthorized clients from using your API quota and accessing your backend resources.

See App Check for Firebase AI Logic for setup instructions. Each platform uses its own attestation provider, so follow the instructions for the target platform.

App Check Debug Tokens for Local Development & CI/CD

Because App Check attestation providers (like Play Integrity or DeviceCheck) reject emulators, simulators, or CI environments, you must use App Check Debug Tokens during development and testing to bypass standard attestation.

[!WARNING] CRITICAL: Never Hardcode or Commit Debug Tokens App Check debug tokens allow clients to bypass attestation and access backend resources without a genuine device. Treat them as private secrets. Never commit debug tokens to version control or hardcode raw token strings in client code (Web, Android, iOS, or Flutter). Always inject them through local environment variables, gitignored local configurations, or CI secrets. If a token is compromised, revoke it immediately in the Firebase Console.

Local Development (Auto-Generated)
  1. Configure your code's App Check provider to use the debug factory. The API is platform-specific; see the App Check section of the platform reference or the App Check documentation linked above.
  2. Run your app in the emulator/localhost.
  3. Look at your runtime debugger console / Logcat logs for the generated UUID:
    • Example: AppCheck debug token: "123a4567-b89c-12d3-e456-789012345678"
  4. Register this token in the Firebase Console under Security > App Check > Apps > Manage debug tokens.

💡 Tip (Prevent Debug Token Churn): Simulator or emulator resets, fresh installs, and clearing browser data can make the SDK generate a new debug token that you must register again. To keep a stable debug token without hardcoding it, see the App Check section of the platform reference.

CI/CD Pipelines (Pre-Provisioned)
  1. Generate and register a new debug token in the Firebase Console under Security > App Check > Apps > Manage debug tokens.
  2. Add this token string as an encrypted secret in your CI system (e.g. APP_CHECK_DEBUG_TOKEN).
  3. Configure your build to pass this secret as an environment variable to the SDK during test execution.

Firebase Remote Config

Consider that you do not need to hardcode model names (e.g., a specific model version string). Use Firebase Remote Config to update model versions dynamically without deploying new client code. See Changing model names remotely


Initialization Code References

[!WARNING] CRITICAL: Use current model names: Always check the Firebase AI Logic Models documentation for the currently supported model names. Do NOT use any Gemini 1.x or Gemini 2.x models (such as gemini-1.5-flash, gemini-2.0-flash, gemini-2.5-flash, or gemini-2.5-flash-lite), as older models are deprecated or shut down.

References

Android (Kotlin) SDK usage patterns iOS SDK code examples and usage patterns Web SDK setup, code examples, and usage patterns

Flutter SDK code examples and usage patterns Client-side Text-to-Speech (TTS) generation

firebaseのその他のスキル

developing-genkit-dart
firebase
Dart向け統合AI SDK。コード生成、構造化出力、ツール、フロー、エージェントを実現。単一インターフェースで生成、ツール定義、フローオーケストレーション、埋め込み、ストリーミングのコアAPIを提供。LLMプロバイダー(Google Gemini、Anthropic Claude、OpenAI GPT)、Firebase AI、Model Context Protocol、Chromeブラウザ統合、Shelf経由のHTTPサーバーホスティング向け8以上のプラグインを含む。フロー実行、トレース、モデル実験などのためのローカル開発UIを備えた組み込みCLI。
developing-genkit-go
firebase
GoでGenkitを使用してAIを活用したアプリケーションを開発します。ユーザーがGenkitを使ってGoでAI機能、エージェント、フロー、ツールを構築するよう依頼した場合や、作業中に使用します。
developing-genkit-js
firebase
Genkitのフロー、ツール、マルチモデルサポートを使用して、AIを活用したNode.js/TypeScriptアプリケーションを構築します。Genkitはプロバイダーに依存せず、プラグインを介してGoogle AI、OpenAI、Anthropic、Ollama、その他のLLMプロバイダーをサポートします。Zodを使用して型安全なスキーマでフローを定義し、生成リクエストを実行し、TypeScriptでマルチステップのAIワークフローを構成します。Genkit CLI v1.29.0以上が必要です。最近の主要なAPI変更により、現在のパターンについては、以前の知識ではなく、genkit docs:readおよびcommon-errors.mdを参照する必要があります...
developing-genkit-python
firebase
PythonでGenkitを使用したAI搭載アプリケーションを開発します。ユーザーがGenkit、AIエージェント、フロー、またはPythonのツールについて質問した場合、またはGenkitに遭遇した場合に使用します…
developing-genkit-tooling
firebase
Genkitツール作成のベストプラクティス。CLIコマンドやMCPサーバーツールを含む。命名規則、アーキテクチャパターン、一貫性について解説…
extension-to-functions-codebase
firebase
インストール済みのFirebase Extension(または拡張機能のソース)を、スタンドアロンのCloud Functions for Firebaseコードベース、または公開可能なnpmパッケージに変換するためのスキル、…
firebase-ai-logic
firebase
クライアントサイドのGemini統合で、マルチモーダル推論、ストリーミング、デバイス上のハイブリッド実行を備えたウェブアプリ向け。テキストのみおよびマルチモーダル入力(画像、音声、動画、PDF)をサポート。20MBを超えるファイルはCloud Storage経由でルーティング。自動履歴付きチャットセッション、リアルタイム表示のためのストリーミング応答、構造化JSON出力の強制を含む。ChromeのGemini Nanoを介したデバイス上のハイブリッド推論を提供し、クラウド実行への自動フォールバックを備える。本番環境ではApp Checkが必要。
firebase-ai-logic
firebase
Firebase AI Logic(Firebase内のGemini)を統合し、インテリジェントなアプリ機能を実現します。FirebaseアプリにAI機能を追加する際や、生成AIを実装する際に使用します。