NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
pub.dev
pub.dev, #2419 most downloaded on pub.dev
Moved to flutter_edge_ai_embeddings. This is the final release under the name flutter_gemma_embeddings; new versions ship as flutter_edge_ai_embeddings, part of Flutter Edge AI.
Latest stable
2.2.2
Last release 1 months ago
05 Oct 2026
Ships fairly regularly
a new release about every 2 weeks
Nearly every release is documented
notes for 11 of 11 stable releases
Nothing withdrawn
no release was ever pulled
5 months old
12 releases · first in 2026
Drag across the ruler to compare, or tap a release to open it.
Moved to flutter_edge_ai_embeddings; this is the final release under the name flutter_gemma_embeddings.
flutter_edge_ai_embeddings; this is the final release under the name flutter_gemma_embeddings.README: engine authors pass activeBackend: to ForwardPassDescriptor.
activeBackend: to ForwardPassDescriptor.Register GemmaEmbeddingTokenizers() via initialize(embeddingTokenizers:) — engines no longer name a tokenizer.
GemmaEmbeddingTokenizers() via initialize(embeddingTokenizers:) — engines no longer name a tokenizer.flutter_gemma_litertlm; re-copy the four JS files from there.web_embedding_model.dart moved to flutter_gemma_litertlm — bump it to 1.8.0.Your coding agent can read these notes before it upgrades. Set up the MCP server →
One column per month.
Lift the dart_sentencepiece_tokenizer cap — 1.4.1 is now required, not excluded.
dart_sentencepiece_tokenizer cap — 1.4.1 is now required, not excluded.Add a SigLIP2 text-embedding profile (no BOS, single trailing EOS, lowercased, fixed 64-token width).
dart:io into the web graph; fromPath throws UnsupportedError there.dart_sentencepiece_tokenizer below 1.4.0, which rejects EmbeddingGemma's and SigLIP2's tokenizer.json.flutter_gemma becomes Flutter Edge AI: every package renamed to flutter_edge_ai* by @DenisovAV in #564
Full Changelog: v1.11.4...v2.0.0
Full Changelog: https://github.com/DenisovAV/flutter_edge_ai/compare/v1.11.4...v2.0.0
LiteRtEmbeddingBackend moved to flutter_gemma_litertlm.ForwardPassDescriptor.tokenizerFactory) so engines can bring WordPiece too.EmbeddingForwardPass.outputContract override + ForwardResult.attentionMask for the mask thread.WordPieceEmbeddingTokenizer (BERT-style, e.g. MiniLM) alongside the existing Gemma SentencePiece adapter.internal: run on the flutter_gemma_litertlm LiteRT engine (no API change).
Migrate to LiteRT v0.14.0 3-arg LiteRtCreateModelFromFile (adds LiteRtEnvironment); native-v0.14.0.
LiteRtCreateModelFromFile (adds LiteRtEnvironment); native-v0.14.0.fix(#318): clamp maxTokens to 1024 + add maxOutputTokens by @DenisovAV in https://github.com/DenisovAV/flutter_gemma/pull/330
Full Changelog: https://github.com/DenisovAV/flutter_gemma/compare/v1.0.1...v1.0.2
1.0.1: restore NPU (#155) + fix empty cancelled chat turn (#325) + README/example fixes by @DenisovAV in https://github.com/DenisovAV/flutter_gemma/pu
Full Changelog: https://github.com/DenisovAV/flutter_gemma/compare/v1.0.0...v1.0.1
homepage to fluttergemma.dev. No code change.chore: pub.dev score 130→150 + WASM compatibility by @DenisovAV in https://github.com/DenisovAV/flutter_gemma/pull/317
Full Changelog: https://github.com/DenisovAV/flutter_gemma/compare/v0.16.5...v1.0.0
dart:io mobile lib for a wasm-clean web graph.Initial release: on-device text embeddings (Gecko / EmbeddingGemma .tflite) via the LiteRT C API + dart:ffi.
.tflite) via the LiteRT C API + dart:ffi.LiteRtEmbeddingBackend (EmbeddingBackendProvider); forward pass runs on a background isolate.