@mediapipe/tasks-retrieval 1.1.0-rc.20260929

package/README.md

@@ -0,0 +1,90 @@
+# MediaPipe Tasks Retrieval Package
+
+This package contains the retrieval tasks for MediaPipe.
+
+## Semantic Retriever
+
+The MediaPipe Semantic Retriever automates vector similarity search and index
+building on the web. It combines in-memory vector storage with embedding-based
+retrieval, allowing you to quickly build multi-modal semantic search into your
+web apps.
+
+For more information, refer to the [Semantic Retriever](https://developers.google.com/mediapipe/solutions/retrieval/semantic_retriever/web_js)
+documentation.
+
+```javascript
+import { SemanticRetriever, InMemoryVectorStore, UniversalEmbedder, FilesetResolver } from "@mediapipe/tasks-retrieval";
+
+const retrieval = await FilesetResolver.forRetrievalTasks("https://cdn.jsdelivr.net/npm/@mediapipe/tasks-retrieval@latest/wasm");
+
+const embedder = await UniversalEmbedder.createFromOptions(retrieval, {
+  baseOptions: {
+    modelAssetPath: "https://storage.googleapis.com/mediapipe-models/universal_embedder/embededder.litertlm"
+  }
+});
+
+const semanticRetriever = await SemanticRetriever.createFromComponents(retrieval, {
+  vectorStore: new InMemoryVectorStore(),
+  providers: [embedder.getProvider()]
+});
+
+await semanticRetriever.insertDocument("doc1", "How to compute semantic distance on the web");
+await semanticRetriever.insertDocument("doc2", "Best vector search libraries for JavaScript");
+
+const results = await semanticRetriever.retrieve("What is vector search in JS?", { limit: 1 });
+console.log(`Similarity Score: ${results[0].score}, Text Data: ${results[0].textData}`);
+```
+
+## Universal Embedder
+
+The MediaPipe Universal Embedder extracts versatile feature embeddings from
+multi-modal inputs – including text, image, and audio - all within the same
+embedding space.
+
+For more information, refer to the [Universal Embedder](https://developers.google.com/mediapipe/solutions/retrieval/universal_embedder/web_js)
+documentation.
+
+```javascript
+import { UniversalEmbedder, FilesetResolver } from "@mediapipe/tasks-retrieval";
+
+const retrieval = await FilesetResolver.forRetrievalTasks("https://cdn.jsdelivr.net/npm/@mediapipe/tasks-retrieval@latest/wasm");
+const universalEmbedder = await UniversalEmbedder.createFromOptions(
+  retrieval,
+  {
+    baseOptions: {
+      modelAssetPath: "https://storage.googleapis.com/mediapipe-models/universal_embedder/embededder.litertlm"
+    }
+  }
+);
+
+// Embed text
+const textEmbeddingResult = universalEmbedder.embedText("The quick brown fox jumps over the lazy dog");
+console.log("Text embedding:", textEmbeddingResult.embeddings[0].floatEmbedding);
+
+// Embed images directly from the DOM
+const image = document.getElementById("myImage");
+const imageEmbeddingResult = universalEmbedder.embedImage(image);
+
+// Compute similarity
+const similarity = UniversalEmbedder.cosineSimilarity(
+  textEmbeddingResult.embeddings[0],
+  imageEmbeddingResult.embeddings[0]
+);
+```
+
+### Privacy Notice
+
+Last modified: June 5, 2026
+
+When you use MediaPipe Tasks, processing of the input data (e.g. images, video,
+text) takes place on device, and MediaPipe does not send that input data to
+Google servers. As a result, you can use our MediaPipe Tasks APIs for
+processing data that should not leave the device.
+
+MediaPipe Tasks APIs send metrics about the performance and utilization of the
+APIs in your app to Google. Google uses this metrics data to measure
+performance, usage, debug, maintain and improve the MediaPipe Tasks, as further
+described in our [Privacy Policy](https://policies.google.com/privacy).
+
+**You are responsible for obtaining informed consent from your app users about
+Google's processing of MediaPipe metrics data as required by applicable law.**

package/package.json

@@ -0,0 +1,29 @@
+{
+  "name": "@mediapipe/tasks-retrieval",
+  "version": "1.1.0-rc.20260929",
+  "description": "MediaPipe Retrieval Tasks. See the privacy notice at https://goo.gle/mediapipe-privacy.",
+  "main": "retrieval_bundle.cjs",
+  "browser": "retrieval_bundle.mjs",
+  "module": "retrieval_bundle.mjs",
+  "exports": {
+    ".": {
+      "import": "./retrieval_bundle.mjs",
+      "require": "./retrieval_bundle.cjs",
+      "default": "./retrieval_bundle.mjs",
+      "types": "./retrieval.d.ts"
+    },
+    "./retrieval_bundle.js": "./retrieval_bundle.js",
+    "./retrieval_wasm_internal.js": "./wasm/retrieval_wasm_internal.js",
+    "./retrieval_wasm_internal.wasm": "./wasm/retrieval_wasm_internal.wasm",
+    "./retrieval_wasm_nosimd_internal.js": "./wasm/retrieval_wasm_nosimd_internal.js",
+    "./retrieval_wasm_nosimd_internal.wasm": "./wasm/retrieval_wasm_nosimd_internal.wasm",
+    "./retrieval_wasm_module_internal.js": "./wasm/retrieval_wasm_module_internal.js",
+    "./retrieval_wasm_module_internal.wasm": "./wasm/retrieval_wasm_module_internal.wasm"
+  },
+  "author": "mediapipe@google.com",
+  "license": "Apache-2.0",
+  "type": "module",
+  "types": "retrieval.d.ts",
+  "homepage": "http://mediapipe.dev",
+  "keywords": [ "AR", "ML", "Augmented", "MediaPipe", "MediaPipe Tasks" ]
+}