Skip to content
pubscout

flutter_edge_ai

2.6.0, released 13 days ago by sashadenisov.dev

On-device AI for Flutter to run Gemma and other LLMs, embeddings, and speech without a server.

WASM-ready
$ flutter pub add flutter_edge_ai
Source repository

What it does

Provides on-device AI capabilities for Flutter apps across Android, iOS, Web, macOS, Windows, and Linux. Supports running various LLMs, vision, audio, embeddings, RAG, and speech using pluggable engines without any cloud or server requirement.

AI-written summary of version 2.6.0, based on the package's README and metadata. Report an inaccuracy

Use it when

  • You need to run large language models or embeddings entirely on-device for privacy or offline use.
  • You want pluggable engines like LiteRT-LM, MediaPipe, ONNX, or llama.cpp within a Flutter application.
  • You require on-device RAG, speech-to-text, or text-to-speech functionality.

Skip it when

  • Your application relies entirely on cloud-hosted APIs and you do not need offline model execution.
  • Targeting platforms or architectures not supported by the required pluggable runtimes.

Strengths

  • Supports multiple pluggable engines (LiteRT-LM, MediaPipe, ONNX, llama.cpp, OS built-in AI).
  • Cross-platform support spanning Android, iOS, Web, macOS, Windows, and Linux.
  • No server or cloud dependency required for running models.

Weaknesses

  • Requires manual platform-specific setup (such as adjusting minimum SDK versions, entitlements, or Podfiles).
  • Managing large model files and runtime binaries increases app complexity and package distribution size.

Setup

Kind
Plugin (native code)
Effort
High
Code generation
No
Direct dependencies
12

Requires platform configurations including minimum SDK versions on Android, Podfile static linkage and entitlements on iOS, copying helper scripts for Web, and macOS deployment target / Podfile setup.

dependencies:
  flutter_edge_ai: ^2.6.0
  flutter_edge_ai_litertlm: ^1.20.0

Alternatives

Packages that do the same job and could replace flutter_edge_ai.

  • firebase_ai

    If you prefer using cloud-hosted Gemini models via Firebase rather than running models locally on the device.

Works with

Packages commonly used together with flutter_edge_ai: ones it builds on, ones built on it, and its own family.