flutter_gemma_rag_qdrant #

qdrant-edge on-device RAG vector store for flutter_gemma. Opt-in package implementing VectorStoreRepository via a Rust FFI shim (qdrant-edge) — ~75× faster search than the legacy sqlite + HNSW path.

Native only (Android, iOS, macOS, Linux, Windows). For web, use flutter_gemma_rag_sqlite (WebSqliteVectorStore).

Usage #

import 'package:flutter_gemma/flutter_gemma.dart';
import 'package:flutter_gemma_rag_qdrant/flutter_gemma_rag_qdrant.dart';

await FlutterGemma.initialize(
  vectorStore: QdrantVectorStore(),
);

Then use the unchanged RAG API:

await FlutterGemmaPlugin.instance.initializeVectorStore('rag_store'); // a directory
await FlutterGemmaPlugin.instance.addDocument(/* ... */);
final hits = await FlutterGemmaPlugin.instance.searchSimilar(query, topK: 5);

QdrantVectorStore also honors the payload-aware Filter DSL on searchSimilar(..., filter: Filter(must: [FieldEquals(key: 'lang', value: 'en')], mustNot: [...])).

The storage path passed to initializeVectorStore is treated as a shard directory (qdrant creates files under it), not a single .db file. Use a distinct path from any sqlite store so they don't collide on disk.

Behavior notes #

• Cross-platform web is not supported — QdrantVectorStore is native-only.
• enableHnsw is accepted but a no-op: qdrant decides indexing internally (brute-forces below ~20k points, which is already faster than the Dart HNSW for typical RAG corpora).
• addDocument's metadata is forwarded as a raw JSON string into the payload; filtering by metadata fields requires valid JSON.
• Distance defaults to cosine.

Platforms #

The native library is fetched at build time by the package's Native-Assets hook (SHA256-verified) — no manual setup.