update lock

Co-authored-by: Marcus Schiesser <mail@marcusschiesser.de>

.changeset/metal-moments-train.md

@@ -0,0 +1,5 @@
+---
+"@llamaindex/community": patch
+---
+
+Added Claude 3.7 Sonnet support

.changeset/selfish-kids-own.md

@@ -0,0 +1,7 @@
+---
+"@llamaindex/core": patch
+"llamaindex": patch
+"@llamaindex/core-tests": patch
+---
+
+feat(node-parser): support async function

.changeset/tender-candles-shop.md

@@ -0,0 +1,5 @@
+---
+"@llamaindex/voyage-ai": major
+---
+
+Adding VoyageAI embedding package

CONTRIBUTING.md

@@ -41,8 +41,15 @@ pnpm install
 
 ### Build the packages
 
+You'll need Turbo to build the packages. If you don't have it, you can run it with `pnpx`.
+
+To build all packages, run:
+
 ```shell
 # Build all packages
+pnpx turbo build --filter "./packages/*"
+
+# Or if you have turbo installed, you can run:
 turbo build --filter "./packages/*"
 ```
 

apps/next/src/content/docs/llamaindex/modules/embeddings/available_embeddings/voyageai.mdx

@@ -0,0 +1,46 @@
+---
+title: VoyageAI
+---
+
+To use VoyageAI embeddings, you need to import `VoyageAIEmbedding` from `@llamaindex/voyage-ai`.
+
+## Installation
+
+import { Tab, Tabs } from "fumadocs-ui/components/tabs";
+
+<Tabs groupId="install" items={["npm", "yarn", "pnpm"]} persist>
+	```shell tab="npm"
+	npm install llamaindex @llamaindex/voyage-ai
+	```
+
+	```shell tab="yarn"
+	yarn add llamaindex @llamaindex/voyage-ai
+	```
+
+	```shell tab="pnpm"
+	pnpm add llamaindex @llamaindex/voyage-ai
+	```
+</Tabs>
+
+```ts
+import { VoyageAIEmbedding } from "@llamaindex/voyage-ai";
+import { Document, Settings, VectorStoreIndex } from "llamaindex";
+
+Settings.embedModel = new VoyageAIEmbedding();
+
+const document = new Document({ text: essay, id_: "essay" });
+
+const index = await VectorStoreIndex.fromDocuments([document]);
+
+const queryEngine = index.asQueryEngine();
+
+const query = "What is the meaning of life?";
+
+const results = await queryEngine.query({
+  query,
+});
+```
+
+## API Reference
+
+- [VoyageAIEmbedding](/docs/api/classes/VoyageAIEmbedding)

apps/next/src/content/docs/llamaindex/modules/embeddings/index.mdx

@@ -37,6 +37,31 @@ Settings.embedModel = new OpenAIEmbedding({
 
 For local embeddings, you can use the [HuggingFace](/docs/llamaindex/modules/embeddings/available_embeddings/huggingface) embedding model.
 
+## Local Ollama Embeddings With Remote Host
+
+Ollama provides a way to run embedding models locally or connect to a remote Ollama instance. This is particularly useful when you need to:
+- Run embeddings without relying on external API services
+- Use custom embedding models
+- Connect to a shared Ollama instance in your network
+
+The ENV variable method you will find elsewhere sometimes may not work with the OllamaEmbedding class. Also note, you'll need to change the host
+in the Ollama server to `0.0.0.0` to allow connections from other machines.
+
+To use Ollama embeddings with a remote host, you need to specify the host URL in the configuration like this:
+
+```typescript
+import { OllamaEmbedding } from "@llamaindex/ollama";
+import { Settings } from "llamaindex";
+
+// Configure Ollama with a remote host
+Settings.embedModel = new OllamaEmbedding({
+  model: "nomic-embed-text",
+  config: {
+    host: "http://your-ollama-host:11434"
+  }
+});
+```
+
 ## Available Embeddings
 
 Most available embeddings are listed in the sidebar on the left.

e2e/examples/nextjs-edge-runtime/package.json

@@ -8,7 +8,7 @@
     "start": "next start"
   },
   "dependencies": {
-    "llamaindex": "workspace:*",
+    "llamaindex": "0.9.3",
     "next": "15.1.7",
     "react": "^18.3.1",
     "react-dom": "^18.3.1"

examples/package.json

@@ -41,6 +41,7 @@
     "@llamaindex/upstash": "^0.0.7",
     "@llamaindex/vercel": "^0.0.13",
     "@llamaindex/vllm": "^0.0.24",
+    "@llamaindex/voyage-ai": "^0.0.1",
     "@llamaindex/weaviate": "^0.0.7",
     "@llamaindex/workflow": "^0.0.11",
     "@notionhq/client": "^2.2.15",

examples/voyage-ai/embedding.ts

@@ -0,0 +1,17 @@
+import { VoyageAIEmbedding } from "@llamaindex/voyage-ai";
+
+async function main() {
+  // API token can be provided as an environment variable too
+  // using VOYAGE_API_TOKEN variable
+  const apiKey = process.env.VOYAGE_API_TOKEN ?? "YOUR_API_TOKEN";
+  const model = "voyage-3-lite";
+  const embedModel = new VoyageAIEmbedding({
+    model,
+    apiKey,
+  });
+  const texts = ["hello", "world"];
+  const embeddings = await embedModel.getTextEmbeddingsBatch(texts);
+  console.log(`\nWe have ${embeddings.length} embeddings`);
+}
+
+main().catch(console.error);

packages/community/src/llm/bedrock/anthropic/types.ts

@@ -6,7 +6,19 @@ export type ToolChoice =
   | { type: "auto" }
   | { type: "tool"; name: string };
 
-export type AnthropicAdditionalChatOptions = { toolChoice: ToolChoice };
+export interface ThinkingConfigDisabled {
+  type: "disabled";
+}
+
+export interface ThinkingConfigEnabled {
+  budget_tokens: number;
+  type: "enabled";
+}
+
+export type AnthropicAdditionalChatOptions = {
+  toolChoice: ToolChoice;
+  thinking?: ThinkingConfigDisabled | ThinkingConfigEnabled;
+};
 
 type Usage = {
   input_tokens: number;

packages/community/src/llm/bedrock/index.ts

@@ -69,6 +69,7 @@ export const BEDROCK_MODELS = {
   ANTHROPIC_CLAUDE_3_5_SONNET: "anthropic.claude-3-5-sonnet-20240620-v1:0",
   ANTHROPIC_CLAUDE_3_5_SONNET_V2: "anthropic.claude-3-5-sonnet-20241022-v2:0",
   ANTHROPIC_CLAUDE_3_5_HAIKU: "anthropic.claude-3-5-haiku-20241022-v1:0",
+  ANTHROPIC_CLAUDE_3_7_SONNET: "anthropic.claude-3-7-sonnet-20250219-v1:0",
   META_LLAMA2_13B_CHAT: "meta.llama2-13b-chat-v1",
   META_LLAMA2_70B_CHAT: "meta.llama2-70b-chat-v1",
   META_LLAMA3_8B_INSTRUCT: "meta.llama3-8b-instruct-v1:0",
@@ -100,6 +101,8 @@ export const INFERENCE_BEDROCK_MODELS = {
     "us.anthropic.claude-3-5-sonnet-20240620-v1:0",
   US_ANTHROPIC_CLAUDE_3_5_SONNET_V2:
     "us.anthropic.claude-3-5-sonnet-20241022-v2:0",
+  US_ANTHROPIC_CLAUDE_3_7_SONNET:
+    "us.anthropic.claude-3-7-sonnet-20250219-v1:0",
   US_META_LLAMA_3_2_1B_INSTRUCT: "us.meta.llama3-2-1b-instruct-v1:0",
   US_META_LLAMA_3_2_3B_INSTRUCT: "us.meta.llama3-2-3b-instruct-v1:0",
   US_META_LLAMA_3_2_11B_INSTRUCT: "us.meta.llama3-2-11b-instruct-v1:0",
@@ -113,6 +116,8 @@ export const INFERENCE_BEDROCK_MODELS = {
   EU_ANTHROPIC_CLAUDE_3_SONNET: "eu.anthropic.claude-3-sonnet-20240229-v1:0",
   EU_ANTHROPIC_CLAUDE_3_5_SONNET:
     "eu.anthropic.claude-3-5-sonnet-20240620-v1:0",
+  EU_ANTHROPIC_CLAUDE_3_7_SONNET:
+    "eu.anthropic.claude-3-7-sonnet-20250219-v1:0",
   EU_META_LLAMA_3_2_1B_INSTRUCT: "eu.meta.llama3-2-1b-instruct-v1:0",
   EU_META_LLAMA_3_2_3B_INSTRUCT: "eu.meta.llama3-2-3b-instruct-v1:0",
 };
@@ -132,6 +137,8 @@ export const INFERENCE_TO_BEDROCK_MAP: Record<
     BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_SONNET,
   [INFERENCE_BEDROCK_MODELS.US_ANTHROPIC_CLAUDE_3_5_SONNET]:
     BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET,
+  [INFERENCE_BEDROCK_MODELS.US_ANTHROPIC_CLAUDE_3_7_SONNET]:
+    BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET,
   [INFERENCE_BEDROCK_MODELS.US_ANTHROPIC_CLAUDE_3_5_SONNET_V2]:
     BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET_V2,
   [INFERENCE_BEDROCK_MODELS.US_ANTHROPIC_CLAUDE_3_5_HAIKU]:
@@ -158,6 +165,8 @@ export const INFERENCE_TO_BEDROCK_MAP: Record<
     BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_SONNET,
   [INFERENCE_BEDROCK_MODELS.EU_ANTHROPIC_CLAUDE_3_5_SONNET]:
     BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET,
+  [INFERENCE_BEDROCK_MODELS.EU_ANTHROPIC_CLAUDE_3_7_SONNET]:
+    BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET,
   [INFERENCE_BEDROCK_MODELS.EU_META_LLAMA_3_2_1B_INSTRUCT]:
     BEDROCK_MODELS.META_LLAMA3_2_1B_INSTRUCT,
   [INFERENCE_BEDROCK_MODELS.EU_META_LLAMA_3_2_3B_INSTRUCT]:
@@ -191,6 +200,7 @@ const CHAT_ONLY_MODELS = {
   [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET]: 200000,
   [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET_V2]: 200000,
   [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_HAIKU]: 200000,
+  [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET]: 200000,
   [BEDROCK_MODELS.META_LLAMA2_13B_CHAT]: 2048,
   [BEDROCK_MODELS.META_LLAMA2_70B_CHAT]: 4096,
   [BEDROCK_MODELS.META_LLAMA3_8B_INSTRUCT]: 8192,
@@ -230,6 +240,7 @@ export const STREAMING_MODELS = new Set([
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET,
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET_V2,
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_HAIKU,
+  BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET,
   BEDROCK_MODELS.META_LLAMA2_13B_CHAT,
   BEDROCK_MODELS.META_LLAMA2_70B_CHAT,
   BEDROCK_MODELS.META_LLAMA3_8B_INSTRUCT,
@@ -256,6 +267,7 @@ export const TOOL_CALL_MODELS: BEDROCK_MODELS[] = [
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET,
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET_V2,
   BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_HAIKU,
+  BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET,
   BEDROCK_MODELS.META_LLAMA3_1_405B_INSTRUCT,
   BEDROCK_MODELS.META_LLAMA3_2_1B_INSTRUCT,
   BEDROCK_MODELS.META_LLAMA3_2_3B_INSTRUCT,
@@ -294,6 +306,7 @@ export const BEDROCK_MODEL_MAX_TOKENS: Partial<Record<BEDROCK_MODELS, number>> =
     [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET]: 4096,
     [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_SONNET_V2]: 8192,
     [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_5_HAIKU]: 8192,
+    [BEDROCK_MODELS.ANTHROPIC_CLAUDE_3_7_SONNET]: 8192,
     [BEDROCK_MODELS.META_LLAMA2_13B_CHAT]: 2048,
     [BEDROCK_MODELS.META_LLAMA2_70B_CHAT]: 2048,
     [BEDROCK_MODELS.META_LLAMA3_8B_INSTRUCT]: 2048,

packages/core/src/node-parser/base.ts

@@ -7,21 +7,27 @@ import {
   TextNode,
   TransformComponent,
 } from "../schema";
+import { isPromise } from "../utils";
 
-export abstract class NodeParser extends TransformComponent<BaseNode[]> {
+export abstract class NodeParser<
+  Result extends TextNode[] | Promise<TextNode[]> =
+    | TextNode[]
+    | Promise<TextNode[]>,
+> extends TransformComponent<Result> {
   includeMetadata: boolean = true;
   includePrevNextRel: boolean = true;
 
   constructor() {
-    super((nodes: BaseNode[]): BaseNode[] => {
+    super((nodes: BaseNode[]): Result => {
+      // alex: should we fix `as` type?
       return this.getNodesFromDocuments(nodes as TextNode[]);
     });
   }
 
   protected postProcessParsedNodes(
-    nodes: TextNode[],
+    nodes: Awaited<Result>,
     parentDocMap: Map<string, TextNode>,
-  ): TextNode[] {
+  ): Awaited<Result> {
     nodes.forEach((node, i) => {
       const parentDoc = parentDocMap.get(node.sourceNode?.nodeId || "");
 
@@ -73,9 +79,9 @@ export abstract class NodeParser extends TransformComponent<BaseNode[]> {
   protected abstract parseNodes(
     documents: TextNode[],
     showProgress?: boolean,
-  ): TextNode[];
+  ): Result;
 
-  public getNodesFromDocuments(documents: TextNode[]): TextNode[] {
+  public getNodesFromDocuments(documents: TextNode[]): Result {
     const docsId: Map<string, TextNode> = new Map(
       documents.map((doc) => [doc.id_, doc]),
     );
@@ -85,20 +91,36 @@ export abstract class NodeParser extends TransformComponent<BaseNode[]> {
       documents,
     });
 
-    const nodes = this.postProcessParsedNodes(
-      this.parseNodes(documents),
-      docsId,
-    );
+    const parsedNodes = this.parseNodes(documents);
+    if (isPromise(parsedNodes)) {
+      return parsedNodes.then((parsedNodes) => {
+        const nodes = this.postProcessParsedNodes(
+          parsedNodes as Awaited<Result>,
+          docsId,
+        );
 
-    callbackManager.dispatchEvent("node-parsing-end", {
-      nodes,
-    });
+        callbackManager.dispatchEvent("node-parsing-end", {
+          nodes,
+        });
 
-    return nodes;
+        return nodes;
+      }) as Result;
+    } else {
+      const nodes = this.postProcessParsedNodes(
+        parsedNodes as Awaited<Result>,
+        docsId,
+      );
+
+      callbackManager.dispatchEvent("node-parsing-end", {
+        nodes,
+      });
+
+      return nodes;
+    }
   }
 }
 
-export abstract class TextSplitter extends NodeParser {
+export abstract class TextSplitter extends NodeParser<TextNode[]> {
   abstract splitText(text: string): string[];
 
   public splitTexts(texts: string[]): string[] {

packages/core/src/node-parser/markdown.ts

@@ -6,7 +6,7 @@ import {
 } from "../schema";
 import { NodeParser } from "./base";
 
-export class MarkdownNodeParser extends NodeParser {
+export class MarkdownNodeParser extends NodeParser<TextNode[]> {
   override parseNodes(nodes: TextNode[], showProgress?: boolean): TextNode[] {
     return nodes.reduce<TextNode[]>((allNodes, node) => {
       const markdownNodes = this.getNodesFromNode(node);

packages/core/src/node-parser/sentence-window.ts

@@ -9,7 +9,7 @@ import {
 import { NodeParser } from "./base";
 import { splitBySentenceTokenizer, type TextSplitterFn } from "./utils";
 
-export class SentenceWindowNodeParser extends NodeParser {
+export class SentenceWindowNodeParser extends NodeParser<TextNode[]> {
   static DEFAULT_WINDOW_SIZE = 3;
   static DEFAULT_WINDOW_METADATA_KEY = "window";
   static DEFAULT_ORIGINAL_TEXT_METADATA_KEY = "originalText";

packages/core/src/utils/index.ts

@@ -1,5 +1,9 @@
 import type { JSONValue } from "../global";
 
+export const isPromise = <T>(obj: unknown): obj is Promise<T> => {
+  return obj != null && typeof obj === "object" && "then" in obj;
+};
+
 export const isAsyncIterable = (
   obj: unknown,
 ): obj is AsyncIterable<unknown> => {

packages/core/tests/node-parser/node-parser.test.ts

@@ -0,0 +1,24 @@
+import { NodeParser } from "@llamaindex/core/node-parser";
+import { TextNode } from "@llamaindex/core/schema";
+import { describe, expect, test } from "vitest";
+
+describe("NodeParser", () => {
+  test("node parser should allow async parse function", async () => {
+    class MyNodeParser extends NodeParser<Promise<TextNode[]>> {
+      protected async parseNodes(documents: TextNode[]): Promise<TextNode[]> {
+        await new Promise((resolve) => setTimeout(resolve, 1000));
+        return documents;
+      }
+    }
+
+    const nodeParser = new MyNodeParser();
+    const nodes = [
+      new TextNode({
+        text: "Hello, world!",
+      }),
+    ];
+    const result = nodeParser(nodes);
+    expect(result).toBeInstanceOf(Promise);
+    await expect(result).resolves.toEqual(nodes);
+  });
+});

packages/llamaindex/src/indices/keyword/index.ts

@@ -296,7 +296,7 @@ export class KeywordTableIndex extends BaseIndex<KeywordTable> {
       await docStore.setDocumentHash(doc.id_, doc.hash);
     }
 
-    const nodes = Settings.nodeParser.getNodesFromDocuments(documents);
+    const nodes = await Settings.nodeParser.getNodesFromDocuments(documents);
     const index = await KeywordTableIndex.init({
       nodes,
       storageContext,

packages/llamaindex/src/indices/summary/index.ts

@@ -145,7 +145,7 @@ export class SummaryIndex extends BaseIndex<IndexList> {
       await docStore.setDocumentHash(doc.id_, doc.hash);
     }
 
-    const nodes = Settings.nodeParser.getNodesFromDocuments(documents);
+    const nodes = await Settings.nodeParser.getNodesFromDocuments(documents);
 
     const index = await SummaryIndex.init({
       nodes,

packages/providers/voyage-ai/package.json

@@ -0,0 +1,40 @@
+{
+  "name": "@llamaindex/voyage-ai",
+  "description": "VoyageAI Adapter for LlamaIndex",
+  "version": "0.0.1",
+  "type": "module",
+  "main": "./dist/index.cjs",
+  "module": "./dist/index.js",
+  "exports": {
+    ".": {
+      "require": {
+        "types": "./dist/index.d.cts",
+        "default": "./dist/index.cjs"
+      },
+      "import": {
+        "types": "./dist/index.d.ts",
+        "default": "./dist/index.js"
+      }
+    }
+  },
+  "files": [
+    "dist"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/run-llama/LlamaIndexTS.git",
+    "directory": "packages/providers/voyage-ai"
+  },
+  "scripts": {
+    "build": "bunchee",
+    "dev": "bunchee --watch"
+  },
+  "devDependencies": {
+    "bunchee": "6.0.3"
+  },
+  "dependencies": {
+    "@llamaindex/core": "workspace:*",
+    "@llamaindex/env": "workspace:*",
+    "voyageai": "0.0.3-1"
+  }
+}

packages/providers/voyage-ai/src/embedding.ts

@@ -0,0 +1,137 @@
+import { BaseEmbedding } from "@llamaindex/core/embeddings";
+import type { MessageContentDetail } from "@llamaindex/core/llms";
+import { extractSingleText } from "@llamaindex/core/utils";
+import { getEnv } from "@llamaindex/env";
+import { VoyageAI, VoyageAIClient } from "voyageai";
+
+const DEFAULT_MODEL = "voyage-3";
+const API_TOKEN_ENV_VARIABLE_NAME = "VOYAGE_API_TOKEN";
+// const API_ROOT = "https://api.voyageai.com/v1/embeddings";
+const DEFAULT_TIMEOUT = 60 * 1000;
+const DEFAULT_MAX_RETRIES = 5;
+
+/**
+ * VoyageAIEmbedding is an alias for VoyageAI that implements the BaseEmbedding interface.
+ */
+export class VoyageAIEmbedding extends BaseEmbedding {
+  /**
+   * VoyageAI model to use
+   * @default "voyage-3"
+   * @see https://docs.voyageai.com/docs/embeddings
+   */
+  model: string;
+
+  /**
+   * VoyageAI API token
+   * @see https://docs.voyageai.com/docs/api-key-and-installation
+   * If not provided, it will try to get the token from the environment variable `VOYAGE_API_KEY`
+   *
+   */
+  apiKey: string;
+
+  /**
+   * Maximum number of retries
+   * @default 5
+   */
+  maxRetries: number;
+
+  /**
+   * Timeout in seconds
+   * @default 60
+   */
+  timeout: number;
+  /**
+   * Whether to truncate the input texts to fit within the context length. Defaults to `true`.
+   * If `true`, over-length input texts will be truncated to fit within the context length, before vectorized by the embedding model.
+   * If `false`, an error will be raised if any given text exceeds the context length.
+   */
+  truncation: boolean;
+
+  /**
+   * VoyageAI supports `document` and `query` as input types, or it can be left undefined. Using an input type prepends the input with a prompt before embedding.
+   * Example from their docs: using "query" adds "Represent the query for retrieving supporting documents:"
+   * VoyageAI says these types improve performance, but it will add to token usage. Embeddings with input types are compatible with those that don't use them.
+   * Setting this to `query` will use the `query` input type for getQueryEmbedding(s).
+   * Setting this to `document` will use the `document` input type for getTextEmbedding(s).
+   * Setting this to `both` will do both of the above.
+   * By default, this is undefined, which means no input types are used.
+   * @see https://docs.voyageai.com/docs/embeddings
+   * @default undefined
+   */
+  useInputTypes: "query" | "document" | "both" | undefined;
+
+  /**
+   * VoyageAI client
+   */
+  client: VoyageAIClient;
+
+  constructor(init?: Partial<VoyageAIEmbedding>) {
+    super();
+
+    this.model = init?.model ?? DEFAULT_MODEL;
+    this.apiKey = init?.apiKey ?? getEnv(API_TOKEN_ENV_VARIABLE_NAME) ?? "";
+    this.maxRetries = init?.maxRetries ?? DEFAULT_MAX_RETRIES;
+    this.timeout = init?.timeout ?? DEFAULT_TIMEOUT;
+    this.truncation = init?.truncation ?? true;
+    this.useInputTypes = init?.useInputTypes;
+    this.client = new VoyageAIClient({
+      apiKey: this.apiKey,
+    });
+  }
+
+  async getTextEmbedding(text: string): Promise<number[]> {
+    const embeddings = await this.getVoyageAIEmbedding([text], "document");
+    return embeddings[0]!;
+  }
+
+  async getQueryEmbedding(
+    query: MessageContentDetail,
+  ): Promise<number[] | null> {
+    const text = extractSingleText(query);
+    if (text) {
+      const embeddings = await this.getVoyageAIEmbedding([text], "query");
+      return embeddings[0]!;
+    } else {
+      return null;
+    }
+  }
+
+  getTextEmbeddings = async (texts: string[]): Promise<number[][]> => {
+    return this.getVoyageAIEmbedding(texts, "document");
+  };
+
+  async getQueryEmbeddings(queries: string[]): Promise<number[][]> {
+    return this.getVoyageAIEmbedding(queries, "query");
+  }
+
+  private getInputType(requestType: "query" | "document") {
+    if (this.useInputTypes === "both") {
+      return requestType;
+    } else if (this.useInputTypes === requestType) {
+      return requestType;
+    } else {
+      return undefined;
+    }
+  }
+
+  private async getVoyageAIEmbedding(
+    inputs: VoyageAI.EmbedRequestInput,
+    inputType: VoyageAI.EmbedRequestInputType,
+  ): Promise<number[][]> {
+    const request: VoyageAI.EmbedRequest = {
+      model: this.model,
+      input: inputs,
+      truncation: this.truncation,
+    };
+    const preferredInputType = this.getInputType(inputType);
+    if (preferredInputType) {
+      request.inputType = preferredInputType;
+    }
+    const response = await this.client.embed(request);
+    if (response.data) {
+      return response.data.map((item) => item.embedding ?? []);
+    } else {
+      throw new Error("Failed to get embeddings from VoyageAI");
+    }
+  }
+}

packages/providers/voyage-ai/src/index.ts

@@ -0,0 +1 @@
+export { VoyageAIEmbedding } from "./embedding";

packages/providers/voyage-ai/tsconfig.json

@@ -0,0 +1,19 @@
+{
+  "extends": "../../../tsconfig.json",
+  "compilerOptions": {
+    "target": "ESNext",
+    "module": "ESNext",
+    "moduleResolution": "bundler",
+    "outDir": "./lib",
+    "tsBuildInfoFile": "./lib/.tsbuildinfo"
+  },
+  "include": ["./src"],
+  "references": [
+    {
+      "path": "../openai/tsconfig.json"
+    },
+    {
+      "path": "../../env/tsconfig.json"
+    }
+  ]
+}

pnpm-lock.yaml

@@ -454,7 +454,7 @@ importers:
   e2e/examples/nextjs-edge-runtime:
     dependencies:
       llamaindex:
-        specifier: workspace:*
+        specifier: 0.9.3
         version: link:../../../packages/llamaindex
       next:
         specifier: 15.1.7
@@ -685,6 +685,9 @@ importers:
       '@llamaindex/vllm':
         specifier: ^0.0.24
         version: link:../packages/providers/vllm
+      '@llamaindex/voyage-ai':
+        specifier: ^0.0.1
+        version: link:../packages/providers/voyage-ai
       '@llamaindex/weaviate':
         specifier: ^0.0.7
         version: link:../packages/providers/storage/weaviate
@@ -1443,7 +1446,7 @@ importers:
         version: link:../../../env
       chromadb:
         specifier: 1.10.3
-        version: 1.10.3(cohere-ai@7.14.0)(openai@4.83.0(ws@8.18.0(bufferutil@4.0.9))(zod@3.24.2))
+        version: 1.10.3(cohere-ai@7.14.0)(openai@4.83.0(ws@8.18.0(bufferutil@4.0.9))(zod@3.24.2))(voyageai@0.0.3-1)
       chromadb-default-embed:
         specifier: ^2.13.2
         version: 2.13.2
@@ -1636,6 +1639,22 @@ importers:
         specifier: 6.3.4
         version: 6.3.4(patch_hash=pavboztthlgni7m5gzw7643oru)(typescript@5.7.3)
 
+  packages/providers/voyage-ai:
+    dependencies:
+      '@llamaindex/core':
+        specifier: workspace:*
+        version: link:../../core
+      '@llamaindex/env':
+        specifier: workspace:*
+        version: link:../../env
+      voyageai:
+        specifier: 0.0.3-1
+        version: 0.0.3-1
+    devDependencies:
+      bunchee:
+        specifier: 6.0.3
+        version: 6.0.3(typescript@5.7.3)
+
   packages/readers:
     dependencies:
       '@azure/cosmos':
@@ -4263,6 +4282,15 @@ packages:
       rollup:
         optional: true
 
+  '@rollup/plugin-node-resolve@15.3.1':
+    resolution: {integrity: sha512-tgg6b91pAybXHJQMAAwW9VuWBO6Thi+q7BCNARLwSqlmsHz0XYURtGvh/AuwSADXSI4h/2uHbs7s4FzlZDGSGA==}
+    engines: {node: '>=14.0.0'}
+    peerDependencies:
+      rollup: ^2.78.0||^3.0.0||^4.0.0
+    peerDependenciesMeta:
+      rollup:
+        optional: true
+
   '@rollup/plugin-node-resolve@16.0.0':
     resolution: {integrity: sha512-0FPvAeVUT/zdWoO0jnb/V5BlBsUSNfkIOtFHzMO4H9MOklrmQFY6FduVHKucNb/aTFxvnGhj4MNj/T1oNdDfNg==}
     engines: {node: '>=14.0.0'}
@@ -5926,6 +5954,16 @@ packages:
     resolution: {integrity: sha512-WDtdLmJvAuNNPzByAYpRo2rF1Mmradw6gvWsQKf63476DDXmomT9zUiGypLcG4ibIM67vhAj8jJRdbmEws2Aqw==}
     engines: {node: '>=6.14.2'}
 
+  bunchee@6.0.3:
+    resolution: {integrity: sha512-Yq/srd3ocXPAHv0KEdJvhFMNUOOVVqy0kNzaGVCirk/+MfnLdvZO5uf5BHugIHe/qSvWUQTJZ3SAfB/VABONeQ==}
+    engines: {node: '>= 18.0.0'}
+    hasBin: true
+    peerDependencies:
+      typescript: ^4.1 || ^5.0
+    peerDependenciesMeta:
+      typescript:
+        optional: true
+
   bunchee@6.3.4:
     resolution: {integrity: sha512-bMy2/+tdMPXOqBAX+9BI0HTNjOXOZ2TXjgFpp5Prt0ztP15xQQUcsECnU7wuBPpLH+4id3rXakH9icdbBRZHZQ==}
     engines: {node: '>= 18.0.0'}
@@ -11534,6 +11572,9 @@ packages:
       jsdom:
         optional: true
 
+  voyageai@0.0.3-1:
+    resolution: {integrity: sha512-R3jN/xnILWoMBL3jPY61Ydm1JbpK3J+VmXBoHvlNg1Xz8h0xdX7sEffXeSu+sAEKQaPyWXVQDmM/jpBhPXw58g==}
+
   vue-demi@0.14.10:
     resolution: {integrity: sha512-nMZBOwuzabUO0nLgIcc6rycZEebF6eeUfaiQx9+WSk8e29IbLvPU9feI6tqW4kTo3hvoYAJkMh8n8D0fuISphg==}
     engines: {node: '>=12'}
@@ -14803,6 +14844,16 @@ snapshots:
     optionalDependencies:
       rollup: 4.34.6
 
+  '@rollup/plugin-node-resolve@15.3.1(rollup@4.34.6)':
+    dependencies:
+      '@rollup/pluginutils': 5.1.4(rollup@4.34.6)
+      '@types/resolve': 1.20.2
+      deepmerge: 4.3.1
+      is-module: 1.0.0
+      resolve: 1.22.10
+    optionalDependencies:
+      rollup: 4.34.6
+
   '@rollup/plugin-node-resolve@16.0.0(rollup@4.34.6)':
     dependencies:
       '@rollup/pluginutils': 5.1.4(rollup@4.34.6)
@@ -16932,6 +16983,31 @@ snapshots:
     dependencies:
       node-gyp-build: 4.8.4
 
+  bunchee@6.0.3(typescript@5.7.3):
+    dependencies:
+      '@rollup/plugin-commonjs': 28.0.2(rollup@4.34.6)
+      '@rollup/plugin-json': 6.1.0(rollup@4.34.6)
+      '@rollup/plugin-node-resolve': 15.3.1(rollup@4.34.6)
+      '@rollup/plugin-replace': 6.0.2(rollup@4.34.6)
+      '@rollup/plugin-wasm': 6.2.2(rollup@4.34.6)
+      '@rollup/pluginutils': 5.1.4(rollup@4.34.6)
+      '@swc/core': 1.10.16(@swc/helpers@0.5.15)
+      '@swc/helpers': 0.5.15
+      clean-css: 5.3.3
+      glob: 11.0.1
+      magic-string: 0.30.17
+      ora: 8.2.0
+      picomatch: 4.0.2
+      pretty-bytes: 5.6.0
+      rollup: 4.34.6
+      rollup-plugin-dts: 6.1.1(rollup@4.34.6)(typescript@5.7.3)
+      rollup-plugin-swc3: 0.11.2(@swc/core@1.10.16(@swc/helpers@0.5.15))(rollup@4.34.6)
+      rollup-preserve-directives: 1.1.3(rollup@4.34.6)
+      tslib: 2.8.1
+      yargs: 17.7.2
+    optionalDependencies:
+      typescript: 5.7.3
+
   bunchee@6.3.4(patch_hash=pavboztthlgni7m5gzw7643oru)(typescript@5.7.2):
     dependencies:
       '@rollup/plugin-commonjs': 28.0.2(rollup@4.34.6)
@@ -17134,13 +17210,14 @@ snapshots:
     transitivePeerDependencies:
       - bare-buffer
 
-  chromadb@1.10.3(cohere-ai@7.14.0)(openai@4.83.0(ws@8.18.0(bufferutil@4.0.9))(zod@3.24.2)):
+  chromadb@1.10.3(cohere-ai@7.14.0)(openai@4.83.0(ws@8.18.0(bufferutil@4.0.9))(zod@3.24.2))(voyageai@0.0.3-1):
     dependencies:
       cliui: 8.0.1
       isomorphic-fetch: 3.0.0
     optionalDependencies:
       cohere-ai: 7.14.0
       openai: 4.83.0(ws@8.18.0(bufferutil@4.0.9))(zod@3.24.2)
+      voyageai: 0.0.3-1
     transitivePeerDependencies:
       - encoding
 
@@ -22738,6 +22815,15 @@ snapshots:
       rollup: 4.34.6
       rollup-preserve-directives: 1.1.3(rollup@4.34.6)
 
+  rollup-plugin-swc3@0.11.2(@swc/core@1.10.16(@swc/helpers@0.5.15))(rollup@4.34.6):
+    dependencies:
+      '@fastify/deepmerge': 1.3.0
+      '@rollup/pluginutils': 5.1.4(rollup@4.34.6)
+      '@swc/core': 1.10.16(@swc/helpers@0.5.15)
+      get-tsconfig: 4.10.0
+      rollup: 4.34.6
+      rollup-preserve-directives: 1.1.3(rollup@4.34.6)
+
   rollup-pluginutils@2.8.2:
     dependencies:
       estree-walker: 0.6.1
@@ -24326,6 +24412,18 @@ snapshots:
       - supports-color
       - terser
 
+  voyageai@0.0.3-1:
+    dependencies:
+      form-data: 4.0.0
+      formdata-node: 6.0.3
+      js-base64: 3.7.2
+      node-fetch: 2.7.0
+      qs: 6.11.2
+      readable-stream: 4.7.0
+      url-join: 4.0.1
+    transitivePeerDependencies:
+      - encoding
+
   vue-demi@0.14.10(vue@3.5.13(typescript@5.7.2)):
     dependencies:
       vue: 3.5.13(typescript@5.7.2)