@aeriondyseti/vector-memory-mcp 0.2.0 → 0.2.1

package/README.md

@@ -179,7 +179,7 @@ vector-memory-mcp/
 
 - **MCP Framework**: @modelcontextprotocol/sdk (official SDK)
 - **Vector Database**: LanceDB (fast, local, vector search)
-- **Embeddings**: @xenova/transformers (Xenova/all-MiniLM-L6-v2, 384 dimensions)
+- **Embeddings**: [@huggingface/transformers](https://huggingface.co/docs/transformers.js) (Xenova/all-MiniLM-L6-v2, 384 dimensions)
 - **Language**: TypeScript 5.0+
 - **Runtime**: Bun 1.0+
 - **Testing**: Bun test
@@ -193,7 +193,7 @@ vector-memory-mcp/
 ```
 Claude Code calls store_memory tool
          ↓
-Content → @xenova/transformers → 384d vector
+Content → @huggingface/transformers → 384d vector
          ↓
 Store in LanceDB with metadata
          ↓
@@ -205,7 +205,7 @@ Store in LanceDB with metadata
 ```
 Claude Code calls search_memories
          ↓
-Query → @xenova/transformers → 384d vector
+Query → @huggingface/transformers → 384d vector
          ↓
 Vector search in LanceDB
          ↓
@@ -346,7 +346,7 @@ MIT License - see [LICENSE](LICENSE) for details.
 
 - Built with [@modelcontextprotocol/sdk](https://github.com/modelcontextprotocol/typescript-sdk) - Official MCP TypeScript SDK
 - Uses [LanceDB](https://lancedb.com/) for fast, local vector search
-- Powered by [@xenova/transformers](https://github.com/xenova/transformers.js) for local embeddings
+- Powered by [@huggingface/transformers](https://huggingface.co/docs/transformers.js) for local embeddings
 - Database layer via [Drizzle ORM](https://orm.drizzle.team/)
 - Inspired by [doobidoo's mcp-memory-service](https://github.com/doobidoo/mcp-memory-service)
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aeriondyseti/vector-memory-mcp",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "A zero-configuration RAG memory server for MCP clients",
   "type": "module",
   "main": "src/index.ts",
@@ -33,7 +33,7 @@
   "dependencies": {
     "@lancedb/lancedb": "^0.22.3",
     "@modelcontextprotocol/sdk": "^1.0.0",
-    "@xenova/transformers": "^2.17.0",
+    "@huggingface/transformers": "^3.8.0",
     "apache-arrow": "^21.1.0"
   },
   "devDependencies": {

package/src/services/embeddings.service.ts

@@ -1,9 +1,9 @@
-import { pipeline, type FeatureExtractionPipeline } from "@xenova/transformers";
+import { pipeline, type FeatureExtractionPipeline } from "@huggingface/transformers";
 
 export class EmbeddingsService {
   private modelName: string;
   private extractor: FeatureExtractionPipeline | null = null;
-  private initPromise: Promise<FeatureExtractionPipeline> | null = null;
+  private initPromise: Promise<void> | null = null;
   private _dimension: number;
 
   constructor(modelName: string, dimension: number) {
@@ -15,32 +15,52 @@ export class EmbeddingsService {
     return this._dimension;
   }
 
-  private async getExtractor(): Promise<FeatureExtractionPipeline> {
+  private async init(): Promise<void> {
     if (this.extractor) {
-      return this.extractor;
+      return; // Already initialized
     }
 
     if (!this.initPromise) {
-      this.initPromise = pipeline("feature-extraction", this.modelName, {
-        quantized: true,
-      }) as Promise<FeatureExtractionPipeline>;
+      this.initPromise = (async () => {
+        this.extractor = await pipeline("feature-extraction", this.modelName, {
+          dtype: "fp32",
+        });
+      })();
     }
-
-    this.extractor = await this.initPromise;
-    return this.extractor;
+    await this.initPromise;
   }
 
   async embed(text: string): Promise<number[]> {
-    const extractor = await this.getExtractor();
-    const output = await extractor(text, { pooling: "mean", normalize: true });
-    return Array.from(output.data as Float32Array);
+    await this.init();
+    if (!this.extractor) {
+      throw new Error("Extractor not initialized.");
+    }
+
+    const output = await this.extractor(text, {
+      pooling: "mean",
+      normalize: true,
+    });
+
+    // output.tolist() returns number[][] for single input, we want the first element
+    const embeddings = output.tolist() as number[][];
+    return embeddings[0];
   }
 
   async embedBatch(texts: string[]): Promise<number[][]> {
-    const results: number[][] = [];
-    for (const text of texts) {
-      results.push(await this.embed(text));
+    await this.init();
+    if (!this.extractor) {
+      throw new Error("Extractor not initialized.");
     }
-    return results;
+
+    if (texts.length === 0) {
+      return [];
+    }
+
+    const output = await this.extractor(texts, {
+      pooling: "mean",
+      normalize: true,
+    });
+
+    return output.tolist() as number[][];
   }
 }