Release 0.11.9 (#2023)

Co-authored-by: marcusschiesser <17126+marcusschiesser@users.noreply.github.com>

apps/next/CHANGELOG.md

@@ -1,5 +1,46 @@
 # @llamaindex/doc
 
+## 0.2.28
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+  - @llamaindex/openai@0.4.5
+  - @llamaindex/cloud@4.0.15
+  - llamaindex@0.11.9
+  - @llamaindex/node-parser@2.0.11
+  - @llamaindex/readers@3.1.9
+  - @llamaindex/workflow@1.1.10
+
+## 0.2.27
+
+### Patch Changes
+
+- 8a51c16: Add natural language agent page
+- Updated dependencies [8a51c16]
+- Updated dependencies [1b5af14]
+  - @llamaindex/workflow@1.1.9
+  - @llamaindex/core@0.6.10
+  - llamaindex@0.11.8
+  - @llamaindex/cloud@4.0.14
+  - @llamaindex/node-parser@2.0.10
+  - @llamaindex/openai@0.4.4
+  - @llamaindex/readers@3.1.8
+
+## 0.2.26
+
+### Patch Changes
+
+- a4d394f: fix: correct SimpleDirectoryReader import path in documentation example
+- Updated dependencies [dbd857f]
+- Updated dependencies [3c857f4]
+  - @llamaindex/workflow@1.1.8
+  - llamaindex@0.11.7
+
 ## 0.2.25
 
 ### Patch Changes

apps/next/next.config.mjs

@@ -15,6 +15,20 @@ const config = {
     "twoslash",
     "typescript",
   ],
+  async redirects() {
+    return [
+      {
+        source: "/docs/chat-ui/:path*.mdx",
+        destination: "/docs/chat-ui/:path*",
+        permanent: true,
+      },
+      {
+        source: "/docs/llamaflow/:path*.mdx",
+        destination: "/docs/llamaflow/:path*",
+        permanent: true,
+      },
+    ];
+  },
   turbopack: {
     resolveAlias: {
       fs: { browser: "./fallback.js" },

apps/next/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/doc",
-  "version": "0.2.25",
+  "version": "0.2.28",
   "private": true,
   "scripts": {
     "postinstall": "fumadocs-mdx",
@@ -16,7 +16,7 @@
     "@huggingface/transformers": "^3.5.0",
     "@icons-pack/react-simple-icons": "^10.1.0",
     "@llama-flow/docs": "0.0.8",
-    "@llamaindex/chat-ui-docs": "0.0.3",
+    "@llamaindex/chat-ui-docs": "^0.0.5",
     "@llamaindex/cloud": "workspace:*",
     "@llamaindex/core": "workspace:*",
     "@llamaindex/node-parser": "workspace:*",
@@ -39,13 +39,13 @@
     "clsx": "2.1.1",
     "foxact": "^0.2.41",
     "framer-motion": "^11.11.17",
-    "fumadocs-core": "^15.2.7",
+    "fumadocs-core": "^15.5.0",
     "fumadocs-docgen": "^2.0.0",
-    "fumadocs-mdx": "^11.6.0",
-    "fumadocs-openapi": "^8.0.1",
-    "fumadocs-twoslash": "^3.1.1",
-    "fumadocs-typescript": "^4.0.2",
-    "fumadocs-ui": "^15.2.7",
+    "fumadocs-mdx": "^11.6.6",
+    "fumadocs-openapi": "^9.0.5",
+    "fumadocs-twoslash": "^3.1.3",
+    "fumadocs-typescript": "^4.0.5",
+    "fumadocs-ui": "^15.5.0",
     "hast-util-to-jsx-runtime": "^2.3.2",
     "llamaindex": "workspace:*",
     "lucide-react": "^0.460.0",

apps/next/source.config.ts

@@ -13,6 +13,11 @@ export const docs = defineDocs({
     "./src/content/docs",
     "./node_modules/@llama-flow/docs",
     "./node_modules/@llamaindex/chat-ui-docs",
+    // NOTE: When adding external docs (like chat-ui or llama-flow above),
+    // make sure to also update:
+    // 1. scripts/validate-links.mts - add to ALLOWED_LINKS array
+    // 2. next.config.mjs - add redirect for .mdx files
+    // 3. src/content/docs/meta.json - add to pages array
   ],
   docs: {
     async: true,

apps/next/src/app/(home)/page.tsx

@@ -113,7 +113,8 @@ export default function HomePage() {
           description="Truly powerful retrieval-augmented generation applications use agentic techniques, and LlamaIndex.TS makes it easy to build them."
         >
           <CodeBlock
-            code={`import { SimpleDirectoryReader, VectorStoreIndex } from "llamaindex";
+            code={`import { VectorStoreIndex } from "llamaindex";
+import { SimpleDirectoryReader } from "@llamaindex/readers/directory";
 import { openai } from "@llamaindex/openai";
 import { agent } from "@llamaindex/workflow";
 

apps/next/src/content/docs/llamaindex/modules/agents/meta.json

@@ -1,4 +1,4 @@
 {
   "title": "Agents",
-  "pages": ["tool", "agent_workflow", "workflows"]
+  "pages": ["tool", "agent_workflow", "workflows", "natural_language_workflow"]
 }

apps/next/src/content/docs/llamaindex/modules/agents/natural_language_workflow.mdx

@@ -0,0 +1,103 @@
+---
+title: Define workflows using natural language  
+---
+
+When working with Workflows, you have to write code to handle an event in the workflow. 
+Often, the logic of the handler is not too complex so that it can be expressed using natural language and executed by an LLM.  
+Besides the instructions, we just need the expected result event of the step, possible tool calls and optionally other events that can be emitted.  
+
+## Usage
+
+Let's take an example of a workflow that generates a joke, gets a critique for it, and then improves it.
+
+### Define the events
+
+First, we define the events for our workflow. We need one for writing the joke, one for critiquing it, and one for the final result:
+
+```typescript
+import { z } from "zod";
+import { zodEvent } from "@llamaindex/workflow";
+
+const writeJokeSchema = z.object({
+  description: z
+    .string()
+    .describe("The topic to write a joke or describe the joke to improve."),
+  writtenJoke: z.optional(z.string()).describe("The written joke."),
+  retriedTimes: z
+    .number()
+    .default(0)
+    .describe(
+      "The retried times for writing the joke. Always increase this from the input retriedTimes.",
+    ),
+});
+
+const critiqueSchema = z.object({
+  joke: z.string().describe("The joke to critique"),
+  retriedTimes: z.number().describe("The retried times for writing the joke."),
+});
+
+const finalResultSchema = z.object({
+  joke: z.string().describe("The joke to critique"),
+  critique: z.string().describe("The critique of the joke"),
+});
+
+const writeJokeEvent = zodEvent(writeJokeSchema, {
+  debugLabel: "writeJokeEvent",
+}); 
+const critiqueEvent = zodEvent(critiqueSchema, {
+  debugLabel: "critiqueEvent",
+}); 
+const finalResultEvent = zodEvent(finalResultSchema, {
+  debugLabel: "finalResultEvent",
+}); 
+```
+
+Note that your natural language workflows the events need to be created by the `zodEvent` function passing the zod schema as an argument. The agent needs the schema of the event data to correctly generate events.  
+Also, we need a `debugLabel` so the LLM can identify the event to emit in the workflow.
+
+### Define the workflow
+
+As usual you first create the workflow:
+
+```typescript
+import { agentHandler, createWorkflow } from "@llamaindex/workflow";
+
+const jokeFlow = createWorkflow();
+```
+
+Then you need to handle the events. For the handlers, instead of code, you're now going to use natural language by calling the `agentHandler` function.
+
+It only requires two parameters:
+- `instructions`: A prompt to guide the agent how to handle the steps.
+- `results`: The output events that the agent should return after handling the step.
+
+Then you will have a simple code to handle the step:
+
+```typescript
+jokeFlow.handle(
+  [writeJokeEvent],
+  agentHandler({
+    instructions: `You are a joke writer. You are given a topic and you need to write a joke about it.`,
+    results: [critiqueEvent],
+  }),
+);
+
+jokeFlow.handle(
+  [critiqueEvent],
+  agentHandler({
+    instructions: `
+You are given a joke and you need to critique it. Follow the following guidelines:
+1. You have maximum 3 times to improve the joke.
+2. If the joke is not good, increase the retriedTimes, describe how to improve the joke and send a writeJokeEvent.
+3. If the joke is good, trigger the finalResultEvent event.
+`,
+    results: [writeJokeEvent, finalResultEvent],
+  }),
+);
+```
+
+For advanced usage, you can add more functionality to `agentHandler` by using these parameters:
+- `events`: A list of additional events that the agent can emit to the workflow. E.g., your agent can emit a `uiEvent` to update the UI during the execution.
+- `tools`: A list of tools that the agent can use to handle the step. E.g., your agent can use a `search` tool to search the web.
+
+You can find more code examples in the [examples](https://github.com/run-llama/LlamaIndexTS/tree/main/examples/agents/natural) folder.

apps/next/src/content/docs/llamaindex/modules/models/llms/openai.mdx

@@ -378,3 +378,186 @@ async function main() {
 ## API Reference
 
 - [OpenAI](/docs/api/classes/OpenAI)
+
+
+# OpenAI Live LLM
+
+The OpenAI Live LLM integration in LlamaIndex provides real-time chat capabilities with support for audio streaming and tool calling.
+
+## Basic Usage
+
+```typescript
+import { openai } from "@llamaindex/openai";
+import { tool, ModalityType } from "llamaindex";
+
+// Get the ephimeral key on the server 
+const serverllm = openai({
+  apiKey: "your-api-key", 
+  model: "gpt-4o-realtime-preview-2025-06-03",
+});
+
+// Get an ephemeral key 
+// Usually this code is run on the server and the ephemeral key is passed to the
+// client - the ephemeral key can be securely used on the client side
+const ephemeralKey = await serverllm.live.getEphemeralKey();
+
+// Create a client-side LLM instance with the ephemeral key
+const llm = openai({
+  apiKey: ephemeralKey,
+  model: "gpt-4o-realtime-preview-2025-06-03"
+});
+
+// Create a live sessionimport { tool } from "llamaindex";
+const session = await llm.live.connect({
+  systemInstruction: "You are a helpful assistant.",
+});
+
+// Send a message
+session.sendMessage({
+  content: "Hello!",
+  role: "user",
+});
+```
+
+## Tool Integration
+
+Tools are handled server-side, making it simple to pass them to the live session:
+
+```typescript
+// Define your tools
+const weatherTool = tool({
+  name: "weather",
+  description: "Get the weather for a location",
+  parameters: z.object({
+    location: z.string().describe("The location to get weather for"),
+  }),
+  execute: async ({ location }) => {
+    return `The weather in ${location} is sunny`;
+  },
+});
+
+// Create session with tools
+const session = await llm.live.connect({
+  systemInstruction: "You are a helpful assistant.",
+  tools: [weatherTool],
+});
+```
+
+## Audio Support
+
+For audio capabilities:
+
+```typescript
+// Get microphone access
+const userStream = await navigator.mediaDevices.getUserMedia({
+  audio: true,
+});
+
+// Create session with audio
+const session = await llm.live.connect({
+  audioConfig: {
+    stream: userStream,
+    onTrack: (remoteStream) => {
+      // Handle incoming audio
+      audioElement.srcObject = remoteStream;
+    },
+  },
+});
+```
+
+## Event Handling
+
+Listen to events from the session:
+
+```typescript
+for await (const event of session.streamEvents()) {
+  if (liveEvents.open.include(event)) {
+    // Connection established
+    console.log("Connected!");
+  } else if (liveEvents.text.include(event)) {
+    // Received text response
+    console.log("Assistant:", event.text);
+  }
+}
+```
+
+## Capabilities
+
+The OpenAI Live LLM supports:
+
+- Real-time text chat
+- Audio streaming (if configured)
+- Tool calling (server-side execution)
+- Ephemeral key generation for secure sessions
+
+## API Reference
+
+### LiveLLM Methods
+// Get an ephemeral key 
+// Usually this code is run on the server and the ephemeral key is passed to the
+// client - the ephemeral key can be securely used on the client side
+
+#### `connect(config?: LiveConnectConfig)`
+
+Creates a new live session.
+
+```typescript
+interface LiveConnectConfig {
+  systemInstruction?: string;
+  tools?: BaseTool[];
+  audioConfig?: AudioConfig;
+  responseModality?: ModalityType[];
+}
+```
+
+#### `getEphemeralKey()`
+
+Gets a temporary key for the session.
+
+### LiveLLMSession Methods
+
+#### `sendMessage(message: ChatMessage)`
+
+Sends a message to the assistant.
+
+```typescript
+interface ChatMessage {
+  content: string | MessageContentDetail[];
+  role: "user" | "assistant";
+}
+```
+
+#### `disconnect()`
+
+Closes the session and cleans up resources.
+
+## Error Handling
+
+```typescript
+try {
+  const session = await llm.live.connect();
+} catch (error) {
+  if (error instanceof Error) {
+    console.error("Connection failed:", error.message);
+  }
+}
+```
+
+## Best Practices
+
+1. **Tool Definition**
+
+   - Keep tool implementations server-side
+   - Use clear descriptions for tools
+   - Handle tool errors gracefully
+
+2. **Session Management**
+
+   - Always disconnect sessions when done
+   - Clean up audio resources
+   - Handle reconnection scenarios
+
+3. **Security**
+   - Use ephemeral keys for sessions
+   - Validate tool inputs
+   - Secure API key handling

e2e/examples/cloudflare-worker-agent/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/cloudflare-worker-agent-test
 
+## 0.0.170
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.0.169
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.0.168
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.0.167
 
 ### Patch Changes

e2e/examples/cloudflare-worker-agent/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/cloudflare-worker-agent-test",
-  "version": "0.0.167",
+  "version": "0.0.170",
   "type": "module",
   "private": true,
   "scripts": {

e2e/examples/llama-parse-browser/CHANGELOG.md

@@ -1,5 +1,17 @@
 # @llamaindex/llama-parse-browser-test
 
+## 0.0.70
+
+### Patch Changes
+
+- @llamaindex/cloud@4.0.15
+
+## 0.0.69
+
+### Patch Changes
+
+- @llamaindex/cloud@4.0.14
+
 ## 0.0.68
 
 ### Patch Changes

e2e/examples/llama-parse-browser/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/llama-parse-browser-test",
   "private": true,
-  "version": "0.0.68",
+  "version": "0.0.70",
   "type": "module",
   "scripts": {
     "dev": "vite",

e2e/examples/nextjs-agent/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/next-agent-test
 
+## 0.1.170
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.1.169
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.1.168
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.1.167
 
 ### Patch Changes

e2e/examples/nextjs-agent/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/next-agent-test",
-  "version": "0.1.167",
+  "version": "0.1.170",
   "private": true,
   "scripts": {
     "dev": "next dev",

e2e/examples/nextjs-edge-runtime/CHANGELOG.md

@@ -1,5 +1,24 @@
 # test-edge-runtime
 
+## 0.1.169
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.1.168
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.1.167
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.1.166
 
 ### Patch Changes

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

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/nextjs-edge-runtime-test",
-  "version": "0.1.166",
+  "version": "0.1.169",
   "private": true,
   "scripts": {
     "dev": "next dev",

e2e/examples/nextjs-node-runtime/CHANGELOG.md

@@ -1,5 +1,28 @@
 # @llamaindex/next-node-runtime
 
+## 0.1.37
+
+### Patch Changes
+
+- llamaindex@0.11.9
+- @llamaindex/huggingface@0.1.15
+- @llamaindex/readers@3.1.9
+
+## 0.1.36
+
+### Patch Changes
+
+- llamaindex@0.11.8
+- @llamaindex/huggingface@0.1.14
+- @llamaindex/readers@3.1.8
+
+## 0.1.35
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.1.34
 
 ### Patch Changes

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

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/next-node-runtime-test",
-  "version": "0.1.34",
+  "version": "0.1.37",
   "private": true,
   "scripts": {
     "dev": "next dev",

e2e/examples/vite-import-llamaindex/CHANGELOG.md

@@ -1,5 +1,24 @@
 # vite-import-llamaindex
 
+## 0.0.36
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.0.35
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.0.34
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.0.33
 
 ### Patch Changes

e2e/examples/vite-import-llamaindex/package.json

@@ -1,7 +1,7 @@
 {
   "name": "vite-import-llamaindex",
   "private": true,
-  "version": "0.0.33",
+  "version": "0.0.36",
   "type": "module",
   "scripts": {
     "build": "vite build",

e2e/examples/vite-import-llamaindex/tsconfig.tsbuildinfo

@@ -1 +1,9 @@
-{"root":["./src/main.ts","./vite.config.ts"],"version":"5.7.3"}
+{
+  "root": [
+    "./src/main.ts",
+    "./vite.config.ts",
+    "./tsconfig.json"
+  ],
+  "errors": true,
+  "version": "5.7.3"
+}

e2e/examples/waku-query-engine/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/waku-query-engine-test
 
+## 0.0.170
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.0.169
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.0.168
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.0.167
 
 ### Patch Changes

e2e/examples/waku-query-engine/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/waku-query-engine-test",
-  "version": "0.0.167",
+  "version": "0.0.170",
   "type": "module",
   "private": true,
   "scripts": {

examples/CHANGELOG.md

@@ -1,5 +1,128 @@
 # examples
 
+## 0.3.23
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+  - @llamaindex/google@0.3.10
+  - @llamaindex/openai@0.4.5
+  - @llamaindex/cloud@4.0.15
+  - llamaindex@0.11.9
+  - @llamaindex/node-parser@2.0.11
+  - @llamaindex/anthropic@0.3.13
+  - @llamaindex/assemblyai@0.1.10
+  - @llamaindex/clip@0.0.61
+  - @llamaindex/cohere@0.0.25
+  - @llamaindex/deepinfra@0.0.61
+  - @llamaindex/discord@0.1.10
+  - @llamaindex/huggingface@0.1.15
+  - @llamaindex/jinaai@0.0.21
+  - @llamaindex/mistral@0.1.11
+  - @llamaindex/mixedbread@0.0.25
+  - @llamaindex/notion@0.1.10
+  - @llamaindex/ollama@0.1.11
+  - @llamaindex/perplexity@0.0.18
+  - @llamaindex/portkey-ai@0.0.53
+  - @llamaindex/replicate@0.0.53
+  - @llamaindex/astra@0.0.25
+  - @llamaindex/azure@0.1.22
+  - @llamaindex/chroma@0.0.25
+  - @llamaindex/elastic-search@0.1.11
+  - @llamaindex/firestore@1.0.18
+  - @llamaindex/milvus@0.1.20
+  - @llamaindex/mongodb@0.0.26
+  - @llamaindex/pinecone@0.1.11
+  - @llamaindex/postgres@0.0.54
+  - @llamaindex/qdrant@0.1.21
+  - @llamaindex/supabase@0.1.10
+  - @llamaindex/upstash@0.0.25
+  - @llamaindex/weaviate@0.0.26
+  - @llamaindex/vercel@0.1.11
+  - @llamaindex/voyage-ai@1.0.17
+  - @llamaindex/readers@3.1.9
+  - @llamaindex/tools@0.0.17
+  - @llamaindex/workflow@1.1.10
+  - @llamaindex/deepseek@0.0.21
+  - @llamaindex/fireworks@0.0.21
+  - @llamaindex/groq@0.0.76
+  - @llamaindex/together@0.0.21
+  - @llamaindex/vllm@0.0.47
+  - @llamaindex/xai@0.0.8
+
+## 0.3.22
+
+### Patch Changes
+
+- Updated dependencies [8a51c16]
+- Updated dependencies [1b5af14]
+  - @llamaindex/workflow@1.1.9
+  - @llamaindex/core@0.6.10
+  - llamaindex@0.11.8
+  - @llamaindex/cloud@4.0.14
+  - @llamaindex/node-parser@2.0.10
+  - @llamaindex/anthropic@0.3.12
+  - @llamaindex/assemblyai@0.1.9
+  - @llamaindex/clip@0.0.60
+  - @llamaindex/cohere@0.0.24
+  - @llamaindex/deepinfra@0.0.60
+  - @llamaindex/discord@0.1.9
+  - @llamaindex/google@0.3.9
+  - @llamaindex/huggingface@0.1.14
+  - @llamaindex/jinaai@0.0.20
+  - @llamaindex/mistral@0.1.10
+  - @llamaindex/mixedbread@0.0.24
+  - @llamaindex/notion@0.1.9
+  - @llamaindex/ollama@0.1.10
+  - @llamaindex/openai@0.4.4
+  - @llamaindex/perplexity@0.0.17
+  - @llamaindex/portkey-ai@0.0.52
+  - @llamaindex/replicate@0.0.52
+  - @llamaindex/astra@0.0.24
+  - @llamaindex/azure@0.1.21
+  - @llamaindex/chroma@0.0.24
+  - @llamaindex/elastic-search@0.1.10
+  - @llamaindex/firestore@1.0.17
+  - @llamaindex/milvus@0.1.19
+  - @llamaindex/mongodb@0.0.25
+  - @llamaindex/pinecone@0.1.10
+  - @llamaindex/postgres@0.0.53
+  - @llamaindex/qdrant@0.1.20
+  - @llamaindex/supabase@0.1.9
+  - @llamaindex/upstash@0.0.24
+  - @llamaindex/weaviate@0.0.25
+  - @llamaindex/vercel@0.1.10
+  - @llamaindex/voyage-ai@1.0.16
+  - @llamaindex/readers@3.1.8
+  - @llamaindex/tools@0.0.16
+  - @llamaindex/deepseek@0.0.20
+  - @llamaindex/fireworks@0.0.20
+  - @llamaindex/groq@0.0.75
+  - @llamaindex/together@0.0.20
+  - @llamaindex/vllm@0.0.46
+  - @llamaindex/xai@0.0.7
+
+## 0.3.21
+
+### Patch Changes
+
+- Updated dependencies [dbd857f]
+- Updated dependencies [3c857f4]
+  - @llamaindex/workflow@1.1.8
+  - llamaindex@0.11.7
+  - @llamaindex/tools@0.0.15
+
+## 0.3.20
+
+### Patch Changes
+
+- Updated dependencies [e7484ef]
+  - @llamaindex/weaviate@0.0.24
+
 ## 0.3.19
 
 ### Patch Changes

examples/agents/agent/blog-writer.ts

@@ -1,4 +1,3 @@
-import { tool } from "@llamaindex/core/tools";
 import { openai } from "@llamaindex/openai";
 import {
   agent,
@@ -7,6 +6,7 @@ import {
   multiAgent,
 } from "@llamaindex/workflow";
 import fs from "fs";
+import { tool } from "llamaindex";
 import os from "os";
 import { z } from "zod";
 

examples/agents/natural/doc-workflow.ts

@@ -0,0 +1,130 @@
+import { ToolCallLLM } from "llamaindex";
+
+import {
+  agentHandler,
+  createWorkflow,
+  workflowEvent,
+  zodEvent,
+} from "@llamaindex/workflow";
+
+import { openai } from "@llamaindex/openai";
+import { z } from "zod";
+
+// ===== 1. Define events =====
+// An event to trigger the workflow
+const planEvent = workflowEvent<{ topic: string }>();
+
+// Generate artifact event
+const ArtifactRequirementSchema = z.object({
+  type: z.literal("markdown"),
+  title: z.string().describe("The title of the artifact."),
+  requirement: z
+    .string()
+    .describe("The requirement for the artifact generation."),
+});
+
+const generateArtifactEvent = zodEvent(ArtifactRequirementSchema, {
+  debugLabel: "generateArtifactEvent",
+});
+
+// Artifact output event
+const ArtifactSchema = z.object({
+  type: z.literal("artifact"),
+  data: z.object({
+    type: z.literal("document"),
+    data: z.object({
+      title: z.string().describe("The title of the data."),
+      content: z.string().describe("The content of the data."),
+      type: z.enum(["markdown", "html"]).describe("The type of the data."),
+    }),
+  }),
+});
+const outputArtifactEvent = zodEvent(ArtifactSchema, {
+  debugLabel: "outputArtifactEvent",
+});
+
+// Events for updating UI
+// assume that we have a UI that can render different states of the workflow
+// and update the UI based on the state and the requirement
+export const UIEventSchema = z.object({
+  type: z.literal("ui_event"),
+  data: z.object({
+    state: z
+      .enum(["plan", "generate", "completed"])
+      .describe("The current state of the workflow."),
+    requirement: z
+      .string()
+      .optional()
+      .describe(
+        "An optional requirement creating or updating a document, if applicable.",
+      ),
+  }),
+});
+const uiEvent = zodEvent(UIEventSchema, { debugLabel: "uiEvent" });
+
+// ===== 2. Define workflow with agents using natural language =====
+// We have a document artifact workflow that made up of 2 steps:
+// 1. Generate requirement for the document
+// 2. Generate document content based on the requirement
+export function createDocumentArtifactWorkflow(llm: ToolCallLLM) {
+  const workflow = createWorkflow();
+
+  // Generate requirement for the document
+  workflow.handle(
+    [planEvent],
+    agentHandler({
+      instructions: `
+Your task is to analyze the request and provide requirements for document generation or update.
+1. Send an uiEvent with the \`plan\` to show UI what you are going to do.
+2. Analyze the conversation history and the user's request carefully to determine the completed tasks and the next steps.
+3. Return the generateArtifactEvent with the requirement for the next step of the document generation or update.
+`,
+      results: [generateArtifactEvent],
+      events: [uiEvent],
+      llm,
+    }),
+  );
+
+  // Generate document content based on the requirement
+  workflow.handle(
+    [generateArtifactEvent],
+    agentHandler({
+      instructions: `
+You are a skilled technical writer who can assist users with documentation.
+Your task is to generate document content based on the requirement and update the UI state.
+
+Here are the steps to handle this task:
+1. First, send an uiEvent with the \`generate\` state and the requirement you received from the input.
+2. Next, start generating the content based on the requirement then send an uiEvent with the \`completed\` state to update the state.
+3. Finally, return the outputArtifactEvent with the document values.
+`,
+      results: [outputArtifactEvent],
+      events: [uiEvent],
+      llm,
+    }),
+  );
+
+  return workflow;
+}
+
+async function main() {
+  const llm = openai({ model: "gpt-4.1-mini" });
+  const workflow = createDocumentArtifactWorkflow(llm);
+  const { stream, sendEvent } = workflow.createContext();
+
+  // Ask the workflow to generate a document about `llama`
+  sendEvent(planEvent.with({ topic: "llama" }));
+
+  await stream.until(outputArtifactEvent).forEach((event) => {
+    if (planEvent.include(event)) {
+      console.log("Starting workflow: ", event.data);
+    }
+    if (uiEvent.include(event)) {
+      console.log("UI event: ", event.data);
+    } else if (outputArtifactEvent.include(event)) {
+      console.log("Output artifact event: ", event.data);
+    }
+  });
+}
+
+main();

examples/agents/natural/joke.ts

@@ -0,0 +1,93 @@
+import { openai } from "@llamaindex/openai";
+import { agentHandler, createWorkflow, zodEvent } from "@llamaindex/workflow";
+import { Settings } from "llamaindex";
+import { z } from "zod";
+
+// Create LLM instance
+const llm = openai({ model: "gpt-4.1-mini" });
+Settings.llm = llm;
+
+// Define our workflow events
+const writeJokeSchema = z.object({
+  description: z
+    .string()
+    .describe("The topic to write a joke or describe the joke to improve."),
+  writtenJoke: z.optional(z.string()).describe("The written joke."),
+  retriedTimes: z
+    .number()
+    .default(0)
+    .describe(
+      "The retried times for writing the joke. Always increase this from the input retriedTimes.",
+    ),
+});
+
+const critiqueSchema = z.object({
+  joke: z.string().describe("The joke to critique"),
+  retriedTimes: z.number().describe("The retried times for writing the joke."),
+});
+
+const finalResultSchema = z.object({
+  joke: z.string().describe("The joke to critique"),
+  critique: z.string().describe("The critique of the joke"),
+});
+
+const writeJokeEvent = zodEvent(writeJokeSchema, {
+  debugLabel: "writeJokeEvent",
+}); // Input topic for writing a joke
+const critiqueEvent = zodEvent(critiqueSchema, {
+  debugLabel: "critiqueEvent",
+}); // Ask for critique of the joke
+const finalResultEvent = zodEvent(finalResultSchema, {
+  debugLabel: "finalResultEvent",
+}); // Final result
+
+// Create our workflow
+const jokeFlow = createWorkflow();
+
+// Define handlers for each step
+// This step always write a joke based on the description
+jokeFlow.handle(
+  [writeJokeEvent],
+  agentHandler({
+    instructions: `You are a joke writer. You are given a topic and you need to write a joke about it.`,
+    results: [critiqueEvent],
+  }),
+);
+
+// This step critiques the joke and asks the writer to improve the joke or send a final result event for stopping.
+jokeFlow.handle(
+  [critiqueEvent],
+  agentHandler({
+    instructions: `
+You are given a joke and you need to critique it. Follow the following guidelines:
+1. You have maximum 3 times to improve the joke.
+2. If the joke is not good, increase the retriedTimes, describe how to improve the joke and send a writeJokeEvent.
+3. If the joke is good, trigger the finalResultEvent event.
+`,
+    results: [writeJokeEvent, finalResultEvent],
+  }),
+);
+
+// Usage
+async function main() {
+  const { stream, sendEvent } = jokeFlow.createContext();
+  sendEvent(writeJokeEvent.with({ description: "write a joke about llama" }));
+
+  await stream.until(finalResultEvent).forEach((event) => {
+    if (writeJokeEvent.include(event)) {
+      console.log(
+        "Triggering write joke: ",
+        JSON.stringify(event.data, null, 2),
+      );
+    } else if (critiqueEvent.include(event)) {
+      console.log("Written joke:  ", JSON.stringify(event.data, null, 2));
+    } else if (finalResultEvent.include(event)) {
+      console.log("Output: ", JSON.stringify(event.data, null, 2));
+    } else {
+      console.log("Unknown event: ", JSON.stringify(event.data, null, 2));
+    }
+  });
+  console.log("Done");
+}
+
+main().catch(console.error);

examples/models/gemini/live/simple.ts

@@ -23,10 +23,7 @@ async function main() {
         content: "Say something about you for 10 seconds",
         role: "user",
       });
-    } else if (
-      liveEvents.audio.include(event) &&
-      typeof event.data === "string"
-    ) {
+    } else if (liveEvents.audio.include(event)) {
       const chunk = Buffer.from(event.data, "base64");
       audioChunks.push(chunk);
       console.log(`Received audio chunk: ${chunk.length} bytes`);

examples/models/gemini/live/tool-call.ts

@@ -1,6 +1,5 @@
-import { ModalityType } from "@llamaindex/core/schema";
-import { tool } from "@llamaindex/core/tools";
 import { gemini, GEMINI_MODEL } from "@llamaindex/google";
+import { ModalityType, tool } from "llamaindex";
 
 import { liveEvents } from "llamaindex";
 import { z } from "zod";

examples/models/openai/live/browser/open-ai-realtime/.gitignore

@@ -0,0 +1,24 @@
+# Logs
+logs
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+pnpm-debug.log*
+lerna-debug.log*
+
+node_modules
+dist
+dist-ssr
+*.local
+
+# Editor directories and files
+.vscode/*
+!.vscode/extensions.json
+.idea
+.DS_Store
+*.suo
+*.ntvs*
+*.njsproj
+*.sln
+*.sw?

examples/models/openai/live/browser/open-ai-realtime/README.md

@@ -0,0 +1,54 @@
+# OpenAI Realtime Chat with LlamaIndex
+
+This is a demo application showcasing real-time audio and text chat capabilities using OpenAI's GPT-4 with voice through LlamaIndex. The application demonstrates bidirectional audio communication and text chat with an AI assistant.
+
+## Features
+
+- Real-time voice communication with GPT-4
+- Text-based chat interface
+- WebRTC-based audio streaming
+- Bidirectional communication (both text and voice)
+- React + TypeScript implementation
+
+## Prerequisites
+
+- Node.js (v18 or higher)
+- OpenAI API key with access to GPT-4 voice models
+- Modern browser with WebRTC support
+
+## Getting Started
+
+1. Install dependencies:
+
+```bash
+pnpm install
+```
+
+2. Start the development server:
+
+```bash
+pnpm run dev
+```
+
+## Usage
+
+The application provides a simple interface where you can:
+
+- Start/Stop a chat session
+- Speak to the AI assistant through your microphone
+- Receive audio responses from the assistant
+- See text transcripts of the conversation
+
+## Technical Details
+
+This project uses:
+
+- LlamaIndex for AI interaction management
+- WebRTC for real-time audio streaming
+- React for the UI
+- Vite for development and building
+- TypeScript for type safety
+
+```
+
+```

examples/models/openai/live/browser/open-ai-realtime/eslint.config.js

@@ -0,0 +1,28 @@
+import js from "@eslint/js";
+import reactHooks from "eslint-plugin-react-hooks";
+import reactRefresh from "eslint-plugin-react-refresh";
+import globals from "globals";
+import tseslint from "typescript-eslint";
+
+export default tseslint.config(
+  { ignores: ["dist"] },
+  {
+    extends: [js.configs.recommended, ...tseslint.configs.recommended],
+    files: ["**/*.{ts,tsx}"],
+    languageOptions: {
+      ecmaVersion: 2020,
+      globals: globals.browser,
+    },
+    plugins: {
+      "react-hooks": reactHooks,
+      "react-refresh": reactRefresh,
+    },
+    rules: {
+      ...reactHooks.configs.recommended.rules,
+      "react-refresh/only-export-components": [
+        "warn",
+        { allowConstantExport: true },
+      ],
+    },
+  },
+);

examples/models/openai/live/browser/open-ai-realtime/index.html

@@ -0,0 +1,13 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <link rel="icon" type="image/svg+xml" href="/vite.svg" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Vite + React + TS</title>
+  </head>
+  <body>
+    <div id="root"></div>
+    <script type="module" src="/src/main.tsx"></script>
+  </body>
+</html>

examples/models/openai/live/browser/open-ai-realtime/package.json

@@ -0,0 +1,29 @@
+{
+  "name": "open-ai-realtime",
+  "private": true,
+  "version": "0.0.0",
+  "type": "module",
+  "scripts": {
+    "dev": "vite",
+    "build": "tsc -b && vite build",
+    "lint": "eslint .",
+    "preview": "vite preview"
+  },
+  "dependencies": {
+    "react": "^19.1.0",
+    "react-dom": "^19.1.0"
+  },
+  "devDependencies": {
+    "@eslint/js": "^9.25.0",
+    "@types/react": "^19.1.2",
+    "@types/react-dom": "^19.1.2",
+    "@vitejs/plugin-react": "^4.5.2",
+    "eslint": "^9.25.0",
+    "eslint-plugin-react-hooks": "^5.2.0",
+    "eslint-plugin-react-refresh": "^0.4.19",
+    "globals": "^16.0.0",
+    "typescript": "~5.8.3",
+    "typescript-eslint": "^8.30.1",
+    "vite": "^6.3.5"
+  }
+}

examples/models/openai/live/browser/open-ai-realtime/public/vite.svg

@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" class="iconify iconify--logos" width="31.88" height="32" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 257"><defs><linearGradient id="IconifyId1813088fe1fbc01fb466" x1="-.828%" x2="57.636%" y1="7.652%" y2="78.411%"><stop offset="0%" stop-color="#41D1FF"></stop><stop offset="100%" stop-color="#BD34FE"></stop></linearGradient><linearGradient id="IconifyId1813088fe1fbc01fb467" x1="43.376%" x2="50.316%" y1="2.242%" y2="89.03%"><stop offset="0%" stop-color="#FFEA83"></stop><stop offset="8.333%" stop-color="#FFDD35"></stop><stop offset="100%" stop-color="#FFA800"></stop></linearGradient></defs><path fill="url(#IconifyId1813088fe1fbc01fb466)" d="M255.153 37.938L134.897 252.976c-2.483 4.44-8.862 4.466-11.382.048L.875 37.958c-2.746-4.814 1.371-10.646 6.827-9.67l120.385 21.517a6.537 6.537 0 0 0 2.322-.004l117.867-21.483c5.438-.991 9.574 4.796 6.877 9.62Z"></path><path fill="url(#IconifyId1813088fe1fbc01fb467)" d="M185.432.063L96.44 17.501a3.268 3.268 0 0 0-2.634 3.014l-5.474 92.456a3.268 3.268 0 0 0 3.997 3.378l24.777-5.718c2.318-.535 4.413 1.507 3.936 3.838l-7.361 36.047c-.495 2.426 1.782 4.5 4.151 3.78l15.304-4.649c2.372-.72 4.652 1.36 4.15 3.788l-11.698 56.621c-.732 3.542 3.979 5.473 5.943 2.437l1.313-2.028l72.516-144.72c1.215-2.423-.88-5.186-3.54-4.672l-25.505 4.922c-2.396.462-4.435-1.77-3.759-4.114l16.646-57.705c.677-2.35-1.37-4.583-3.769-4.113Z"></path></svg>

examples/models/openai/live/browser/open-ai-realtime/src/audio-chat.tsx

@@ -0,0 +1,183 @@
+import { openai } from "@llamaindex/openai";
+import { liveEvents, LiveLLMSession, ModalityType } from "llamaindex";
+import { useEffect, useRef, useState } from "react";
+
+const MicIcon = ({ isConnected }: { isConnected: boolean }) => (
+  <svg
+    xmlns="http://www.w3.org/2000/svg"
+    viewBox="0 0 24 24"
+    fill="currentColor"
+    strokeWidth="2"
+    strokeLinecap="round"
+    strokeLinejoin="round"
+  >
+    {isConnected ? (
+      <>
+        <path d="M12 1a3 3 0 0 0-3 3v8a3 3 0 0 0 6 0V4a3 3 0 0 0-3-3z" />
+        <path d="M19 10v2a7 7 0 0 1-14 0v-2" />
+        <line x1="12" y1="19" x2="12" y2="23" />
+        <line x1="8" y1="23" x2="16" y2="23" />
+      </>
+    ) : (
+      <>
+        <path d="M12 1a3 3 0 0 0-3 3v8a3 3 0 0 0 6 0V4a3 3 0 0 0-3-3z" />
+        <path d="M19 10v2a7 7 0 0 1-14 0v-2" />
+      </>
+    )}
+  </svg>
+);
+
+const WaveAnimation = () => (
+  <div className="wave-animation">
+    {[...Array(3)].map((_, i) => (
+      <div key={i} className="wave" style={{ animationDelay: `${i * 0.2}s` }} />
+    ))}
+  </div>
+);
+
+export const AudioChat = () => {
+  const [isConnected, setIsConnected] = useState(false);
+  const [messages, setMessages] = useState<
+    Array<{ role: string; content: string }>
+  >([]);
+  const [status, setStatus] = useState<string>("");
+  const audioRef = useRef<HTMLAudioElement>(null);
+  const sessionRef = useRef<LiveLLMSession | null>(null);
+  const [stream, setStream] = useState<MediaStream | null>(null);
+  const messagesEndRef = useRef<HTMLDivElement>(null);
+
+  const scrollToBottom = () => {
+    messagesEndRef.current?.scrollIntoView({ behavior: "smooth" });
+  };
+
+  useEffect(() => {
+    scrollToBottom();
+  }, [messages]);
+
+  useEffect(() => {
+    return () => {
+      if (stream) {
+        stream.getTracks().forEach((track) => track.stop());
+      }
+    };
+  }, [stream]);
+
+  const startChat = async () => {
+    try {
+      setStatus("Initializing microphone...");
+      const userStream = await navigator.mediaDevices.getUserMedia({
+        audio: true,
+      });
+      setStream(userStream);
+
+      setStatus("Connecting to AI...");
+      const apiKey = prompt("Please enter your OpenAI API key:");
+      if (!apiKey) {
+        throw new Error("API key is required");
+      }
+
+      // move this call to the server side for security reasons
+      // Do not store the API key in the frontend!
+      const serverllm = openai({
+        apiKey: apiKey,
+        model: "gpt-4o-realtime-preview-2025-06-03",
+      });
+
+      const tempKey = await serverllm.live.getEphemeralKey();
+
+      const llm = openai({
+        apiKey: tempKey,
+        model: "gpt-4o-realtime-preview-2025-06-03",
+      });
+      const session = await llm.live.connect({
+        systemInstruction: "You are a helpful assistant who speaks naturally.",
+        responseModality: [ModalityType.TEXT, ModalityType.AUDIO],
+        audioConfig: {
+          stream: userStream,
+          onTrack: (remoteStream) => {
+            if (audioRef.current && remoteStream) {
+              audioRef.current.srcObject = remoteStream;
+              audioRef.current.play().catch(console.error);
+            }
+          },
+        },
+      });
+
+      sessionRef.current = session;
+      setIsConnected(true);
+      setStatus("Connected! Listening...");
+
+      for await (const event of session.streamEvents()) {
+        if (liveEvents.open.include(event)) {
+          setMessages((prev) => [
+            ...prev,
+            {
+              role: "user",
+              content: "Hello, I'm ready to chat!",
+            },
+          ]);
+          session.sendMessage({
+            content: "Hello, I'm ready to chat!",
+            role: "user",
+          });
+        } else if (liveEvents.text.include(event)) {
+          setMessages((prev) => [
+            ...prev,
+            {
+              role: "assistant",
+              content: event.text,
+            },
+          ]);
+        }
+      }
+    } catch (error) {
+      console.error("Error starting chat:", error);
+      setStatus("Error connecting. Please try again.");
+      setIsConnected(false);
+    }
+  };
+
+  const stopChat = async () => {
+    setStatus("Disconnecting...");
+    if (sessionRef.current) {
+      await sessionRef.current.disconnect();
+      sessionRef.current = null;
+    }
+    if (stream) {
+      stream.getTracks().forEach((track) => track.stop());
+      setStream(null);
+    }
+    if (audioRef.current) {
+      audioRef.current.srcObject = null;
+    }
+    setIsConnected(false);
+    setStatus("");
+  };
+
+  return (
+    <div className="audio-chat-container">
+      <h1>AI Voice Chat</h1>
+      <div className="messages-container">
+        {messages.map((msg, idx) => (
+          <div key={idx} className={`message ${msg.role}`}>
+            {msg.content}
+          </div>
+        ))}
+        <div ref={messagesEndRef} />
+      </div>
+
+      <div className="controls">
+        {status && <div className="status-indicator">{status}</div>}
+        <button
+          className={`mic-button ${isConnected ? "connected" : ""}`}
+          onClick={isConnected ? stopChat : startChat}
+          title={isConnected ? "Stop Chat" : "Start Chat"}
+        >
+          <MicIcon isConnected={isConnected} />
+          {isConnected && <WaveAnimation />}
+        </button>
+        <audio ref={audioRef} style={{ display: "none" }} />
+      </div>
+    </div>
+  );
+};

examples/models/openai/live/browser/open-ai-realtime/src/index.css

@@ -0,0 +1,322 @@
+:root {
+  --primary-color: #646cff;
+  --secondary-color: #535bf2;
+  --background-dark: #1a1a1a;
+  --chat-bg: #242424;
+  --text-primary: #ffffff;
+  --text-secondary: #888888;
+  --success-color: #4caf50;
+  --error-color: #f44336;
+  --gradient-start: #4776e6;
+  --gradient-end: #8e54e9;
+}
+
+body {
+  background-color: var(--background-dark);
+  color: var(--text-primary);
+  margin: 0;
+  min-height: 100vh;
+  font-family:
+    -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Oxygen, Ubuntu,
+    Cantarell, "Open Sans", "Helvetica Neue", sans-serif;
+}
+
+#root {
+  max-width: 1280px;
+  height: 100vh;
+  margin: 0 auto;
+  padding: 2rem;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+}
+
+.logo {
+  height: 6em;
+  padding: 1.5em;
+  will-change: filter;
+  transition: filter 300ms;
+}
+.logo:hover {
+  filter: drop-shadow(0 0 2em #646cffaa);
+}
+.logo.react:hover {
+  filter: drop-shadow(0 0 2em #61dafbaa);
+}
+
+@keyframes logo-spin {
+  from {
+    transform: rotate(0deg);
+  }
+  to {
+    transform: rotate(360deg);
+  }
+}
+
+@media (prefers-reduced-motion: no-preference) {
+  a:nth-of-type(2) .logo {
+    animation: logo-spin infinite 20s linear;
+  }
+}
+
+.card {
+  padding: 2em;
+}
+
+.read-the-docs {
+  color: #888;
+}
+
+.audio-chat-container {
+  display: flex;
+  flex-direction: column;
+  gap: 2rem;
+  width: 100%;
+  max-width: 800px;
+  height: 80vh;
+  margin: 0 auto;
+  padding: 2rem;
+  background: var(--chat-bg);
+  border-radius: 24px;
+  box-shadow: 0 8px 32px rgba(0, 0, 0, 0.2);
+  position: relative;
+  overflow: hidden;
+}
+
+.audio-chat-container::before {
+  content: "";
+  position: absolute;
+  top: 0;
+  left: 0;
+  right: 0;
+  height: 4px;
+  background: linear-gradient(
+    to right,
+    var(--gradient-start),
+    var(--gradient-end)
+  );
+}
+
+.audio-chat-container h1 {
+  font-size: 2.5rem;
+  margin: 0;
+  background: linear-gradient(
+    to right,
+    var(--gradient-start),
+    var(--gradient-end)
+  );
+  -webkit-background-clip: text;
+  background-clip: text;
+  color: transparent;
+  text-align: center;
+}
+
+.messages-container {
+  display: flex;
+  flex-direction: column;
+  gap: 1rem;
+  flex: 1;
+  overflow-y: auto;
+  padding: 1rem;
+  border-radius: 16px;
+  background: rgba(255, 255, 255, 0.05);
+  backdrop-filter: blur(10px);
+  margin: 1rem 0;
+}
+
+.message {
+  padding: 1rem 1.5rem;
+  border-radius: 16px;
+  max-width: 80%;
+  text-align: left;
+  animation: messageSlide 0.3s ease-out;
+  box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
+}
+
+@keyframes messageSlide {
+  from {
+    opacity: 0;
+    transform: translateY(20px);
+  }
+  to {
+    opacity: 1;
+    transform: translateY(0);
+  }
+}
+
+.message.user {
+  background: linear-gradient(
+    135deg,
+    var(--gradient-start),
+    var(--gradient-end)
+  );
+  align-self: flex-end;
+  margin-left: 20%;
+  color: white;
+}
+
+.message.assistant {
+  background: rgba(255, 255, 255, 0.1);
+  align-self: flex-start;
+  margin-right: 20%;
+  border: 1px solid rgba(255, 255, 255, 0.1);
+}
+
+.controls {
+  display: flex;
+  justify-content: center;
+  align-items: center;
+  padding: 2rem;
+  position: relative;
+}
+
+.mic-button {
+  width: 80px;
+  height: 80px;
+  border-radius: 50%;
+  border: none;
+  background: linear-gradient(
+    135deg,
+    var(--gradient-start),
+    var(--gradient-end)
+  );
+  color: white;
+  cursor: pointer;
+  transition: all 0.3s cubic-bezier(0.4, 0, 0.2, 1);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  box-shadow: 0 4px 15px rgba(0, 0, 0, 0.2);
+  position: relative;
+  overflow: hidden;
+}
+
+.mic-button::before {
+  content: "";
+  position: absolute;
+  top: 0;
+  left: 0;
+  right: 0;
+  bottom: 0;
+  background: linear-gradient(
+    135deg,
+    rgba(255, 255, 255, 0.1),
+    rgba(255, 255, 255, 0)
+  );
+  border-radius: 50%;
+  transition: transform 0.3s ease;
+}
+
+.mic-button:hover {
+  transform: scale(1.05);
+  box-shadow: 0 8px 25px rgba(0, 0, 0, 0.3);
+}
+
+.mic-button:hover::before {
+  transform: translateY(-100%);
+}
+
+.mic-button.connected {
+  background: var(--error-color);
+  animation: pulseError 2s infinite;
+}
+
+.mic-button svg {
+  width: 32px;
+  height: 32px;
+  filter: drop-shadow(0 2px 4px rgba(0, 0, 0, 0.2));
+  transition: transform 0.3s ease;
+}
+
+.mic-button:hover svg {
+  transform: scale(1.1);
+}
+
+@keyframes pulseError {
+  0% {
+    box-shadow: 0 0 0 0 rgba(244, 67, 54, 0.4);
+  }
+  70% {
+    box-shadow: 0 0 0 20px rgba(244, 67, 54, 0);
+  }
+  100% {
+    box-shadow: 0 0 0 0 rgba(244, 67, 54, 0);
+  }
+}
+
+/* Status indicator */
+.status-indicator {
+  position: absolute;
+  top: -30px;
+  left: 50%;
+  transform: translateX(-50%);
+  font-size: 0.9rem;
+  color: var(--text-secondary);
+  opacity: 0;
+  transition: opacity 0.3s ease;
+}
+
+.controls:hover .status-indicator {
+  opacity: 1;
+}
+
+/* Scrollbar styling */
+.messages-container::-webkit-scrollbar {
+  width: 8px;
+}
+
+.messages-container::-webkit-scrollbar-track {
+  background: rgba(255, 255, 255, 0.05);
+  border-radius: 4px;
+}
+
+.messages-container::-webkit-scrollbar-thumb {
+  background: linear-gradient(var(--gradient-start), var(--gradient-end));
+  border-radius: 4px;
+}
+
+.messages-container::-webkit-scrollbar-thumb:hover {
+  background: linear-gradient(var(--gradient-end), var(--gradient-start));
+}
+
+/* Wave Animation */
+.wave-animation {
+  position: absolute;
+  bottom: -15px;
+  left: 50%;
+  transform: translateX(-50%);
+  display: flex;
+  gap: 4px;
+}
+
+.wave {
+  width: 4px;
+  height: 15px;
+  background: currentColor;
+  border-radius: 2px;
+  animation: wave 0.5s ease-in-out infinite;
+}
+
+@keyframes wave {
+  0%,
+  100% {
+    transform: scaleY(0.5);
+  }
+  50% {
+    transform: scaleY(1.5);
+  }
+}
+
+/* Loading state */
+.mic-button.loading {
+  animation: rotate 1s linear infinite;
+}
+
+@keyframes rotate {
+  from {
+    transform: rotate(0deg);
+  }
+  to {
+    transform: rotate(360deg);
+  }
+}

examples/models/openai/live/browser/open-ai-realtime/src/main.tsx

@@ -0,0 +1,10 @@
+import { StrictMode } from "react";
+import { createRoot } from "react-dom/client";
+import { AudioChat } from "./audio-chat.tsx";
+import "./index.css";
+
+createRoot(document.getElementById("root")!).render(
+  <StrictMode>
+    <AudioChat />
+  </StrictMode>,
+);

examples/models/openai/live/browser/open-ai-realtime/src/vite-env.d.ts

@@ -0,0 +1 @@
+/// <reference types="vite/client" />

examples/models/openai/live/browser/open-ai-realtime/tsconfig.app.json

@@ -0,0 +1,26 @@
+{
+  "compilerOptions": {
+    "tsBuildInfoFile": "./node_modules/.tmp/tsconfig.app.tsbuildinfo",
+    "target": "ES2020",
+    "useDefineForClassFields": true,
+    "lib": ["ES2020", "DOM", "DOM.Iterable"],
+    "module": "ESNext",
+    "skipLibCheck": true,
+
+    /* Bundler mode */
+    "moduleResolution": "bundler",
+    "allowImportingTsExtensions": true,
+    "verbatimModuleSyntax": true,
+    "moduleDetection": "force",
+    "noEmit": true,
+    "jsx": "react-jsx",
+
+    /* Linting */
+    "strict": true,
+    "noUnusedLocals": true,
+    "noUnusedParameters": true,
+    "noFallthroughCasesInSwitch": true,
+    "noUncheckedSideEffectImports": true
+  },
+  "include": ["src"]
+}

examples/models/openai/live/browser/open-ai-realtime/tsconfig.json

@@ -0,0 +1,4 @@
+{
+  "files": [],
+  "references": [{ "path": "./tsconfig.app.json" }]
+}

examples/models/openai/live/browser/open-ai-realtime/vite.config.ts

@@ -0,0 +1,7 @@
+import react from "@vitejs/plugin-react";
+import { defineConfig } from "vite";
+
+// https://vite.dev/config/
+export default defineConfig({
+  plugins: [react()],
+});

examples/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/examples",
-  "version": "0.3.19",
+  "version": "0.3.23",
   "private": true,
   "scripts": {
     "lint": "eslint .",
@@ -11,51 +11,51 @@
     "@azure/cosmos": "^4.1.1",
     "@azure/identity": "^4.4.1",
     "@azure/search-documents": "^12.1.0",
-    "@llamaindex/anthropic": "^0.3.11",
-    "@llamaindex/assemblyai": "^0.1.8",
-    "@llamaindex/astra": "^0.0.23",
-    "@llamaindex/azure": "^0.1.20",
-    "@llamaindex/chroma": "^0.0.23",
-    "@llamaindex/clip": "^0.0.59",
-    "@llamaindex/cloud": "^4.0.13",
-    "@llamaindex/cohere": "^0.0.23",
-    "@llamaindex/core": "^0.6.9",
-    "@llamaindex/deepinfra": "^0.0.59",
-    "@llamaindex/deepseek": "^0.0.19",
-    "@llamaindex/discord": "^0.1.8",
-    "@llamaindex/elastic-search": "^0.1.9",
+    "@llamaindex/anthropic": "^0.3.13",
+    "@llamaindex/assemblyai": "^0.1.10",
+    "@llamaindex/astra": "^0.0.25",
+    "@llamaindex/azure": "^0.1.22",
+    "@llamaindex/chroma": "^0.0.25",
+    "@llamaindex/clip": "^0.0.61",
+    "@llamaindex/cloud": "^4.0.15",
+    "@llamaindex/cohere": "^0.0.25",
+    "@llamaindex/core": "^0.6.11",
+    "@llamaindex/deepinfra": "^0.0.61",
+    "@llamaindex/deepseek": "^0.0.21",
+    "@llamaindex/discord": "^0.1.10",
+    "@llamaindex/elastic-search": "^0.1.11",
     "@llamaindex/env": "^0.1.30",
-    "@llamaindex/firestore": "^1.0.16",
-    "@llamaindex/fireworks": "^0.0.19",
-    "@llamaindex/google": "^0.3.6",
-    "@llamaindex/groq": "^0.0.74",
-    "@llamaindex/huggingface": "^0.1.13",
-    "@llamaindex/jinaai": "^0.0.19",
-    "@llamaindex/milvus": "^0.1.18",
-    "@llamaindex/mistral": "^0.1.9",
-    "@llamaindex/mixedbread": "^0.0.23",
-    "@llamaindex/mongodb": "^0.0.24",
-    "@llamaindex/node-parser": "^2.0.9",
-    "@llamaindex/notion": "^0.1.8",
-    "@llamaindex/ollama": "^0.1.9",
-    "@llamaindex/openai": "^0.4.3",
-    "@llamaindex/perplexity": "^0.0.16",
-    "@llamaindex/pinecone": "^0.1.9",
-    "@llamaindex/portkey-ai": "^0.0.51",
-    "@llamaindex/postgres": "^0.0.52",
-    "@llamaindex/qdrant": "^0.1.19",
-    "@llamaindex/readers": "^3.1.7",
-    "@llamaindex/replicate": "^0.0.51",
-    "@llamaindex/supabase": "^0.1.8",
-    "@llamaindex/together": "^0.0.19",
-    "@llamaindex/tools": "^0.0.14",
-    "@llamaindex/upstash": "^0.0.23",
-    "@llamaindex/vercel": "^0.1.9",
-    "@llamaindex/vllm": "^0.0.45",
-    "@llamaindex/voyage-ai": "^1.0.15",
-    "@llamaindex/weaviate": "^0.0.23",
-    "@llamaindex/workflow": "^1.1.6",
-    "@llamaindex/xai": "workspace:^0.0.6",
+    "@llamaindex/firestore": "^1.0.18",
+    "@llamaindex/fireworks": "^0.0.21",
+    "@llamaindex/google": "^0.3.10",
+    "@llamaindex/groq": "^0.0.76",
+    "@llamaindex/huggingface": "^0.1.15",
+    "@llamaindex/jinaai": "^0.0.21",
+    "@llamaindex/milvus": "^0.1.20",
+    "@llamaindex/mistral": "^0.1.11",
+    "@llamaindex/mixedbread": "^0.0.25",
+    "@llamaindex/mongodb": "^0.0.26",
+    "@llamaindex/node-parser": "^2.0.11",
+    "@llamaindex/notion": "^0.1.10",
+    "@llamaindex/ollama": "^0.1.11",
+    "@llamaindex/openai": "^0.4.5",
+    "@llamaindex/perplexity": "^0.0.18",
+    "@llamaindex/pinecone": "^0.1.11",
+    "@llamaindex/portkey-ai": "^0.0.53",
+    "@llamaindex/postgres": "^0.0.54",
+    "@llamaindex/qdrant": "^0.1.21",
+    "@llamaindex/readers": "^3.1.9",
+    "@llamaindex/replicate": "^0.0.53",
+    "@llamaindex/supabase": "^0.1.10",
+    "@llamaindex/together": "^0.0.21",
+    "@llamaindex/tools": "^0.0.17",
+    "@llamaindex/upstash": "^0.0.25",
+    "@llamaindex/vercel": "^0.1.11",
+    "@llamaindex/vllm": "^0.0.47",
+    "@llamaindex/voyage-ai": "^1.0.17",
+    "@llamaindex/weaviate": "^0.0.26",
+    "@llamaindex/workflow": "^1.1.10",
+    "@llamaindex/xai": "workspace:^0.0.8",
     "@notionhq/client": "^2.2.15",
     "@pinecone-database/pinecone": "^4.0.0",
     "@vercel/postgres": "^0.10.0",
@@ -64,7 +64,7 @@
     "commander": "^12.1.0",
     "dotenv": "^16.4.5",
     "js-tiktoken": "^1.0.14",
-    "llamaindex": "^0.11.5",
+    "llamaindex": "^0.11.9",
     "mongodb": "6.7.0",
     "postgres": "^3.4.4",
     "wikipedia": "^2.1.2",

examples/rag/longText.ts

@@ -1,3 +1,4 @@
+import { OpenAIEmbedding } from "@llamaindex/openai";
 import {
   Document,
   SentenceSplitter,
@@ -7,6 +8,10 @@ import {
 import { OldSentenceSplitter } from "./old-sentence-splitter";
 export const STORAGE_DIR = "./data";
 
+Settings.embedModel = new OpenAIEmbedding({
+  model: "text-embedding-3-small",
+});
+
 // Update node parser
 (async () => {
   // generate a document with a very long sentence (9000 words long)

examples/tsconfig.json

@@ -18,5 +18,11 @@
       "module": "commonjs"
     }
   },
-  "include": ["./**/*.ts"]
+  "include": ["./**/*.ts"],
+  "exclude": [
+    "node_modules",
+    "dist",
+    "models/openai/live/browser/open-ai-realtime",
+    "**/browser/**"
+  ]
 }

packages/autotool/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/autotool
 
+## 8.0.9
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 8.0.8
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 8.0.7
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 8.0.6
 
 ### Patch Changes

packages/autotool/examples/01_node/CHANGELOG.md

@@ -1,5 +1,27 @@
 # @llamaindex/autotool-01-node-example
 
+## 0.0.117
+
+### Patch Changes
+
+- llamaindex@0.11.9
+- @llamaindex/autotool@8.0.9
+
+## 0.0.116
+
+### Patch Changes
+
+- llamaindex@0.11.8
+- @llamaindex/autotool@8.0.8
+
+## 0.0.115
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+  - @llamaindex/autotool@8.0.7
+
 ## 0.0.114
 
 ### Patch Changes

packages/autotool/examples/01_node/package.json

@@ -13,5 +13,5 @@
   "scripts": {
     "start": "node --import tsx --import @llamaindex/autotool/node ./src/index.ts"
   },
-  "version": "0.0.114"
+  "version": "0.0.117"
 }

packages/autotool/package.json

@@ -6,7 +6,7 @@
     "url": "git+https://github.com/run-llama/LlamaIndexTS.git",
     "directory": "packages/autotool"
   },
-  "version": "8.0.6",
+  "version": "8.0.9",
   "description": "auto transpile your JS function to LLM Agent compatible",
   "files": [
     "dist",

packages/cloud/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/cloud
 
+## 4.0.15
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 4.0.14
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 4.0.13
 
 ### Patch Changes

packages/cloud/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/cloud",
-  "version": "4.0.13",
+  "version": "4.0.15",
   "type": "module",
   "license": "MIT",
   "scripts": {

packages/core/CHANGELOG.md

@@ -1,5 +1,20 @@
 # @llamaindex/core
 
+## 0.6.11
+
+### Patch Changes
+
+- a89e187: Feat: added custom abbreviations to sentence splitter
+- 62699b7: Improve performance of sentence splitter
+- c5b2691: Add more Acronyms on SentenceSplitter
+- d8ac8d3: Feat: add support for openai realtime API
+
+## 0.6.10
+
+### Patch Changes
+
+- 1b5af14: fix: jsonToNode for image nodes
+
 ## 0.6.9
 
 ### Patch Changes

packages/core/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/core",
   "type": "module",
-  "version": "0.6.9",
+  "version": "0.6.11",
   "description": "LlamaIndex Core Module",
   "exports": {
     "./agent": {

packages/core/src/llms/base.ts

@@ -1,4 +1,5 @@
-import { extractText, streamConverter } from "../utils";
+import { extractText } from "../utils/llms";
+import { streamConverter } from "../utils/stream";
 import type {
   ChatResponse,
   ChatResponseChunk,

packages/core/src/llms/index.ts

@@ -1,6 +1,9 @@
 export { BaseLLM, ToolCallLLM } from "./base";
-export { LiveLLM, LiveLLMSession, liveEvents, type LiveEvent } from "./live";
+export { LiveLLM, LiveLLMCapability, LiveLLMSession } from "./live/live";
+export { liveEvents, type LiveEvent } from "./live/live-types";
+export type { MessageSender } from "./live/sender";
 export type {
+  AudioConfig,
   BaseTool,
   BaseToolWithCall,
   ChatMessage,

packages/core/src/llms/live/live-types.ts

@@ -1,9 +1,7 @@
 import type {
-  ChatMessage,
-  LiveConnectConfig,
   MessageContentAudioDetail,
   MessageContentTextDetail,
-} from "./type";
+} from "../type";
 
 export type OpenEvent = { type: "open" };
 
@@ -63,45 +61,3 @@ export const liveEvents = {
       e.type === "turnComplete",
   },
 };
-export abstract class LiveLLMSession {
-  protected eventQueue: LiveEvent[] = [];
-  protected eventResolvers: ((value: LiveEvent) => void)[] = [];
-  protected closed = false;
-  abstract sendMessage(message: ChatMessage): void;
-  async *streamEvents(): AsyncIterable<LiveEvent> {
-    while (true) {
-      const event = await this.nextEvent();
-      if (event === undefined) {
-        break;
-      }
-      yield event;
-    }
-  }
-  abstract disconnect(): Promise<void>;
-
-  protected async nextEvent(): Promise<LiveEvent | undefined> {
-    if (this.eventQueue.length) {
-      return Promise.resolve(this.eventQueue.shift());
-    }
-
-    return new Promise((resolve) => {
-      this.eventResolvers.push(resolve);
-    });
-  }
-
-  //Uses an async queue to send events to the client
-  // if the consumer is waiting for an event, it will be resolved immediately
-  // otherwise, the event will be queued up and sent when the consumer is ready
-  pushEventToQueue(event: LiveEvent) {
-    if (this.eventResolvers.length) {
-      //resolving the promise with the event
-      this.eventResolvers.shift()!(event);
-    } else {
-      this.eventQueue.push(event);
-    }
-  }
-}
-
-export abstract class LiveLLM {
-  abstract connect(config?: LiveConnectConfig): Promise<LiveLLMSession>;
-}

packages/core/src/llms/live/live.ts

@@ -0,0 +1,124 @@
+import type {
+  ChatMessage,
+  LiveConnectConfig,
+  MessageContentAudioDetail,
+  MessageContentDetail,
+  MessageContentImageDataDetail,
+  MessageContentVideoDetail,
+} from "../type";
+import type { LiveEvent } from "./live-types";
+import type { MessageSender } from "./sender";
+
+export enum LiveLLMCapability {
+  EPHEMERAL_KEY = "ephemeral_key",
+  AUDIO_CONFIG = "audio_config",
+}
+
+export abstract class LiveLLMSession {
+  protected eventQueue: LiveEvent[] = [];
+  protected eventResolvers: ((value: LiveEvent) => void)[] = [];
+  closed = false;
+  abstract get messageSender(): MessageSender;
+
+  private isTextMessage(content: MessageContentDetail) {
+    return content.type === "text";
+  }
+
+  private isAudioMessage(
+    content: MessageContentDetail,
+  ): content is MessageContentAudioDetail {
+    return content.type === "audio";
+  }
+
+  private isImageMessage(
+    content: MessageContentDetail,
+  ): content is MessageContentImageDataDetail {
+    return content.type === "image";
+  }
+
+  private isVideoMessage(
+    content: MessageContentDetail,
+  ): content is MessageContentVideoDetail {
+    return content.type === "video";
+  }
+
+  sendMessage(message: ChatMessage) {
+    const { content, role } = message;
+    if (!Array.isArray(content)) {
+      this.messageSender.sendTextMessage(content, role);
+    } else {
+      for (const item of content) {
+        this.processMessage(item, role);
+      }
+    }
+  }
+
+  private processMessage(message: MessageContentDetail, role?: string) {
+    if (this.isTextMessage(message)) {
+      this.messageSender.sendTextMessage(message.text, role);
+    } else if (
+      this.isAudioMessage(message) &&
+      this.messageSender.sendAudioMessage
+    ) {
+      this.messageSender.sendAudioMessage(message, role);
+    } else if (
+      this.isImageMessage(message) &&
+      this.messageSender.sendImageMessage
+    ) {
+      this.messageSender.sendImageMessage(message, role);
+    } else if (
+      this.isVideoMessage(message) &&
+      this.messageSender.sendVideoMessage
+    ) {
+      this.messageSender.sendVideoMessage(message, role);
+    }
+  }
+
+  async *streamEvents(): AsyncIterable<LiveEvent> {
+    while (true) {
+      const event = await this.nextEvent();
+      if (event === undefined) {
+        break;
+      }
+      yield event;
+    }
+  }
+  abstract disconnect(): Promise<void>;
+
+  protected async nextEvent(): Promise<LiveEvent | undefined> {
+    if (this.eventQueue.length) {
+      return Promise.resolve(this.eventQueue.shift());
+    }
+
+    return new Promise((resolve) => {
+      this.eventResolvers.push(resolve);
+    });
+  }
+
+  //Uses an async queue to send events to the client
+  // if the consumer is waiting for an event, it will be resolved immediately
+  // otherwise, the event will be queued up and sent when the consumer is ready
+  pushEventToQueue(event: LiveEvent) {
+    if (this.eventResolvers.length) {
+      //resolving the promise with the event
+      this.eventResolvers.shift()!(event);
+    } else {
+      this.eventQueue.push(event);
+    }
+  }
+}
+
+export abstract class LiveLLM {
+  /**
+   * Set of capabilities supported by this implementation.
+   * Override in subclasses as needed.
+   */
+  capabilities: Set<LiveLLMCapability> = new Set();
+
+  abstract connect(config?: LiveConnectConfig): Promise<LiveLLMSession>;
+  abstract getEphemeralKey(): Promise<string | undefined>;
+
+  hasCapability(capability: LiveLLMCapability): boolean {
+    return this.capabilities.has(capability);
+  }
+}

packages/core/src/llms/live/sender.ts

@@ -0,0 +1,15 @@
+import type {
+  MessageContentAudioDetail,
+  MessageContentImageDataDetail,
+  MessageContentVideoDetail,
+} from "../type";
+
+export interface MessageSender {
+  sendTextMessage(message: string, role?: string): void;
+  sendAudioMessage?(content: MessageContentAudioDetail, role?: string): void;
+  sendImageMessage?(
+    content: MessageContentImageDataDetail,
+    role?: string,
+  ): void;
+  sendVideoMessage?(content: MessageContentVideoDetail, role?: string): void;
+}

packages/core/src/llms/type.ts

@@ -290,8 +290,14 @@ export type ToolOutput = {
   isError: boolean;
 };
 
+export interface AudioConfig {
+  stream?: MediaStream;
+  onTrack?: (track: MediaStream | null) => void;
+}
+
 export interface LiveConnectConfig {
   tools?: BaseTool[];
   responseModality?: ModalityType[];
   systemInstruction?: string;
+  audioConfig?: AudioConfig;
 }

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

@@ -0,0 +1,42 @@
+export const abbreviations = {
+  english: [
+    "i.e.",
+    "etc.",
+    "vs.",
+    "Inc.",
+    "A.S.A.P.",
+    "Mr.",
+    "Mrs.",
+    "Ms.",
+    "Dr.",
+    "Prof.",
+    "Sr.",
+    "Jr.",
+    "Tel.",
+    "a.m.",
+    "p.m.",
+    "Art.",
+  ],
+  spanish: [
+    "Sr.",
+    "Sres.",
+    "Srs.",
+    "Sra.",
+    "Sras.",
+    "Srta.",
+    "Srtas.",
+    "Dr.",
+    "Drs.",
+    "Dra.",
+    "Dras.",
+    "Prof.",
+    "Profs.",
+    "Profa.",
+    "Profas.",
+    "Ing.",
+    "Lic.",
+    "Arq.",
+    "Ab.",
+    "Abs.",
+  ],
+};

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

@@ -12,11 +12,7 @@ import {
   type TextSplitterFn,
 } from "./utils";
 
-type _Split = {
-  text: string;
-  isSentence: boolean;
-  tokenSize: number;
-};
+type _Split = [string, boolean, number]; // [text, isSentence, tokenSize]
 
 /**
  * Parse text with a preference for complete sentences.
@@ -42,6 +38,11 @@ export class SentenceSplitter extends MetadataAwareTextSplitter {
    * Backup regex for splitting into sentences.
    */
   secondaryChunkingRegex: string = "[^,.;。？！]+[,.;。？！]?";
+  /**
+   * Extra abbreviations to consider while splitting into sentences.
+   * For example, for contracts, you may want to consider "LLC." as an important abbreviation
+   */
+  extraAbbreviations: string[] | undefined = [];
 
   #chunkingTokenizerFn = splitBySentenceTokenizer();
   #splitFns: Set<TextSplitterFn> = new Set();
@@ -59,7 +60,11 @@ export class SentenceSplitter extends MetadataAwareTextSplitter {
       this.separator = parsedParams.separator;
       this.paragraphSeparator = parsedParams.paragraphSeparator;
       this.secondaryChunkingRegex = parsedParams.secondaryChunkingRegex;
+      this.extraAbbreviations = parsedParams.extraAbbreviations;
     }
+    this.#chunkingTokenizerFn = splitBySentenceTokenizer(
+      this.extraAbbreviations,
+    );
     this.#tokenizer = params?.tokenizer ?? Settings.tokenizer;
     this.#splitFns.add(splitBySep(this.paragraphSeparator));
     this.#splitFns.add(this.#chunkingTokenizerFn);
@@ -105,34 +110,26 @@ export class SentenceSplitter extends MetadataAwareTextSplitter {
     return chunks;
   }
 
-  #split(text: string, chunkSize: number): _Split[] {
-    const tokenSize = this.tokenSize(text);
-    if (tokenSize <= chunkSize) {
-      return [
-        {
-          text,
-          isSentence: true,
-          tokenSize,
-        },
-      ];
+  *#split(
+    text: string,
+    chunkSize: number,
+    tokenSize?: number,
+  ): Generator<_Split> {
+    const _tokenSize = tokenSize ?? this.tokenSize(text);
+    if (_tokenSize <= chunkSize) {
+      yield [text, true, _tokenSize];
+      return;
     }
     const [textSplitsByFns, isSentence] = this.#getSplitsByFns(text);
-    const textSplits: _Split[] = [];
 
     for (const textSplit of textSplitsByFns) {
-      const tokenSize = this.tokenSize(textSplit);
-      if (tokenSize <= chunkSize) {
-        textSplits.push({
-          text: textSplit,
-          isSentence,
-          tokenSize,
-        });
+      const textSplitTokenSize = this.tokenSize(textSplit);
+      if (textSplitTokenSize <= chunkSize) {
+        yield [textSplit, isSentence, textSplitTokenSize];
       } else {
-        const recursiveTextSplits = this.#split(textSplit, chunkSize);
-        textSplits.push(...recursiveTextSplits);
+        yield* this.#split(textSplit, chunkSize, textSplitTokenSize);
       }
     }
-    return textSplits;
   }
 
   #getSplitsByFns(text: string): [splits: string[], isSentence: boolean] {
@@ -151,7 +148,7 @@ export class SentenceSplitter extends MetadataAwareTextSplitter {
     return [[text], true];
   }
 
-  #merge(splits: _Split[], chunkSize: number): string[] {
+  #merge(splits: Iterable<_Split>, chunkSize: number): string[] {
     const chunks: string[] = [];
     let currentChunk: [string, number][] = [];
     let lastChunk: [string, number][] = [];
@@ -177,23 +174,26 @@ export class SentenceSplitter extends MetadataAwareTextSplitter {
       }
     };
 
-    while (splits.length > 0) {
-      const curSplit = splits[0]!;
-      if (curSplit.tokenSize > chunkSize) {
+    const splitsIterator = splits[Symbol.iterator]();
+    let currentSplitResult = splitsIterator.next();
+
+    while (!currentSplitResult.done) {
+      const [text, isSentence, tokenSize] = currentSplitResult.value;
+      if (tokenSize > chunkSize) {
         throw new Error("Single token exceeded chunk size");
       }
-      if (currentChunkLength + curSplit.tokenSize > chunkSize && !newChunk) {
+      if (currentChunkLength + tokenSize > chunkSize && !newChunk) {
         closeChunk();
       } else {
         if (
-          curSplit.isSentence ||
-          currentChunkLength + curSplit.tokenSize <= chunkSize ||
+          isSentence ||
+          currentChunkLength + tokenSize <= chunkSize ||
           newChunk
         ) {
-          currentChunkLength += curSplit.tokenSize;
-          currentChunk.push([curSplit.text, curSplit.tokenSize]);
-          splits.shift();
+          currentChunkLength += tokenSize;
+          currentChunk.push([text, tokenSize]);
           newChunk = false;
+          currentSplitResult = splitsIterator.next();
         } else {
           closeChunk();
         }

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

@@ -1,3 +1,4 @@
+import { abbreviations } from "./abbreviations";
 import type { TextSplitter } from "./base";
 import SentenceTokenizer from "./sentence_tokenizer";
 
@@ -34,19 +35,15 @@ export const splitByChar = (): TextSplitterFn => {
   return (text: string) => text.split("");
 };
 
-let sentenceTokenizer: SentenceTokenizer | null = null;
-
-export const splitBySentenceTokenizer = (): TextSplitterFn => {
-  if (!sentenceTokenizer) {
-    sentenceTokenizer = new SentenceTokenizer([
-      "i.e.",
-      "etc.",
-      "vs.",
-      "Inc.",
-      "A.S.A.P.",
-    ]);
-  }
-  const tokenizer = sentenceTokenizer;
+export const splitBySentenceTokenizer = (
+  extraAbbreviations: string[] | undefined = [],
+): TextSplitterFn => {
+  const tokenizer = new SentenceTokenizer([
+    ...abbreviations.english,
+    ...abbreviations.spanish,
+    // Add the extra abbreviations provided by the user, e.g. for business-specific context
+    ...extraAbbreviations,
+  ]);
   return (text: string) => {
     try {
       return tokenizer.tokenize(text);

packages/core/src/schema/node.ts

@@ -347,6 +347,8 @@ export function jsonToNode(json: any, type?: ObjectType) {
       return new Document(json);
     case ObjectType.IMAGE_DOCUMENT:
       return new ImageDocument(json);
+    case ObjectType.IMAGE:
+      return new ImageNode(json);
     default:
       throw new Error(`Invalid node type: ${nodeType}`);
   }

packages/core/src/schema/zod.ts

@@ -51,6 +51,13 @@ export const sentenceSplitterSchema = z
       })
       .optional()
       .default("[^,.;。？！]+[,.;。？！]?"),
+    extraAbbreviations: z
+      .array(z.string(), {
+        description:
+          "Extra abbreviations to consider while splitting into sentences. For example, for contracts, you may want to consider 'LLC.' as an important abbreviation",
+      })
+      .optional()
+      .default([]),
   })
   .refine(
     (data) => data.chunkOverlap < data.chunkSize,

packages/core/src/utils/index.ts

@@ -14,19 +14,6 @@ export const isIterable = (obj: unknown): obj is Iterable<unknown> => {
   return obj != null && typeof obj === "object" && Symbol.iterator in obj;
 };
 
-export async function* streamConverter<S, D>(
-  stream: AsyncIterable<S>,
-  converter: (s: S) => D | null,
-): AsyncIterable<D> {
-  for await (const data of stream) {
-    const newData = converter(data);
-    if (newData === null) {
-      return;
-    }
-    yield newData;
-  }
-}
-
 export async function* streamCallbacks<S>(
   stream: AsyncIterable<S>,
   callbacks: {
@@ -86,3 +73,4 @@ export {
 export { MockLLM } from "./mock";
 
 export { objectEntries } from "./object-entries";
+export * from "./stream";

packages/core/src/utils/stream.ts

@@ -0,0 +1,12 @@
+export async function* streamConverter<S, D>(
+  stream: AsyncIterable<S>,
+  converter: (s: S) => D | null,
+): AsyncIterable<D> {
+  for await (const data of stream) {
+    const newData = converter(data);
+    if (newData === null) {
+      return;
+    }
+    yield newData;
+  }
+}

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

@@ -75,6 +75,22 @@ describe("sentence splitter", () => {
     expect(splits).toEqual(["This is a sentence. This is another sentence."]);
   });
 
+  test("overall split long text", () => {
+    const sentenceSplitter = new SentenceSplitter({
+      chunkSize: 10,
+      chunkOverlap: 0,
+    });
+    const splits = sentenceSplitter.splitText(
+      "The first short sentence. The first long long long sentence. The second short sentence. The second long long long sentence.",
+    );
+    expect(splits).toEqual([
+      "The first short sentence.",
+      "The first long long long sentence.",
+      "The second short sentence.",
+      "The second long long long sentence.",
+    ]);
+  });
+
   test("doesn't split decimals", () => {
     const sentenceSplitter = new SentenceSplitter({
       chunkSize: 5,
@@ -91,6 +107,39 @@ describe("sentence splitter", () => {
     ]);
   });
 
+  test("doesn't split basic abbreviations", () => {
+    const sentenceSplitter = new SentenceSplitter({
+      chunkSize: 15,
+      chunkOverlap: 0,
+    });
+    const splits = sentenceSplitter.splitText(
+      "This is a sentence of Broda Noel. This is the sentence of Sr. Broda Noel. This is a sentence of somebody else",
+    );
+
+    expect(splits).toEqual([
+      "This is a sentence of Broda Noel.",
+      "This is the sentence of Sr. Broda Noel.",
+      "This is a sentence of somebody else",
+    ]);
+  });
+
+  test("doesn't split extra abbreviations", () => {
+    const sentenceSplitter = new SentenceSplitter({
+      chunkSize: 10,
+      chunkOverlap: 0,
+      extraAbbreviations: ["S.A."],
+    });
+    const splits = sentenceSplitter.splitText(
+      "This is a sentence. The S.A. Broda Company. This is another sentence",
+    );
+
+    expect(splits).toEqual([
+      "This is a sentence.",
+      "The S.A. Broda Company.",
+      "This is another sentence",
+    ]);
+  });
+
   test("splits cjk", () => {
     const sentenceSplitter = new SentenceSplitter({
       chunkSize: 30,

packages/experimental/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/experimental
 
+## 0.0.186
+
+### Patch Changes
+
+- llamaindex@0.11.9
+
+## 0.0.185
+
+### Patch Changes
+
+- llamaindex@0.11.8
+
+## 0.0.184
+
+### Patch Changes
+
+- Updated dependencies [3c857f4]
+  - llamaindex@0.11.7
+
 ## 0.0.183
 
 ### Patch Changes

packages/experimental/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/experimental",
   "description": "Experimental package for LlamaIndexTS",
-  "version": "0.0.183",
+  "version": "0.0.186",
   "type": "module",
   "types": "dist/type/index.d.ts",
   "main": "dist/cjs/index.js",

packages/llamaindex/CHANGELOG.md

@@ -1,5 +1,37 @@
 # llamaindex
 
+## 0.11.9
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+  - @llamaindex/cloud@4.0.15
+  - @llamaindex/node-parser@2.0.11
+  - @llamaindex/workflow@1.1.10
+
+## 0.11.8
+
+### Patch Changes
+
+- Updated dependencies [8a51c16]
+- Updated dependencies [1b5af14]
+  - @llamaindex/workflow@1.1.9
+  - @llamaindex/core@0.6.10
+  - @llamaindex/cloud@4.0.14
+  - @llamaindex/node-parser@2.0.10
+
+## 0.11.7
+
+### Patch Changes
+
+- 3c857f4: chore: move ajv to dev deps
+- Updated dependencies [dbd857f]
+  - @llamaindex/workflow@1.1.8
+
 ## 0.11.6
 
 ### Patch Changes

packages/llamaindex/package.json

@@ -1,6 +1,6 @@
 {
   "name": "llamaindex",
-  "version": "0.11.6",
+  "version": "0.11.9",
   "license": "MIT",
   "type": "module",
   "keywords": [
@@ -24,14 +24,14 @@
     "@llamaindex/core": "workspace:*",
     "@llamaindex/env": "workspace:*",
     "@llamaindex/node-parser": "workspace:*",
-    "@llamaindex/workflow": "1.1.7",
+    "@llamaindex/workflow": "workspace:*",
     "@types/lodash": "^4.17.7",
     "@types/node": "^22.9.0",
-    "ajv": "^8.17.1",
     "lodash": "^4.17.21",
     "magic-bytes.js": "^1.10.0"
   },
   "devDependencies": {
+    "ajv": "^8.17.1",
     "@types/node": "^22.9.0"
   },
   "engines": {

packages/llamaindex/tests/CHANGELOG.md

@@ -1,5 +1,18 @@
 # @llamaindex/core-test
 
+## 0.1.6
+
+### Patch Changes
+
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/openai@0.4.5
+
+## 0.1.5
+
+### Patch Changes
+
+- @llamaindex/openai@0.4.4
+
 ## 0.1.4
 
 ### Patch Changes

packages/llamaindex/tests/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/llamaindex-test",
   "private": true,
-  "version": "0.1.4",
+  "version": "0.1.6",
   "type": "module",
   "scripts": {
     "test": "vitest run"

packages/node-parser/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/node-parser
 
+## 2.0.11
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 2.0.10
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 2.0.9
 
 ### Patch Changes

packages/node-parser/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@llamaindex/node-parser",
-  "version": "2.0.9",
+  "version": "2.0.11",
   "description": "Node parser for LlamaIndex",
   "type": "module",
   "exports": {

packages/providers/anthropic/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/anthropic
 
+## 0.3.13
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.3.12
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 0.3.11
 
 ### Patch Changes

packages/providers/anthropic/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/anthropic",
   "description": "Anthropic Adapter for LlamaIndex",
-  "version": "0.3.11",
+  "version": "0.3.13",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/assemblyai/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/assemblyai
 
+## 0.1.10
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.1.9
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 0.1.8
 
 ### Patch Changes

packages/providers/assemblyai/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/assemblyai",
   "description": "AssemblyAI Reader for LlamaIndex",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "type": "module",
   "types": "dist/index.d.ts",
   "main": "dist/index.cjs",

packages/providers/aws/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/community
 
+## 0.0.106
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.0.105
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 0.0.104
 
 ### Patch Changes

packages/providers/aws/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/aws",
   "description": "AWS package for LlamaIndexTS",
-  "version": "0.0.104",
+  "version": "0.0.106",
   "type": "module",
   "types": "dist/type/index.d.ts",
   "main": "dist/cjs/index.js",

packages/providers/clip/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/clip
 
+## 0.0.61
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+  - @llamaindex/openai@0.4.5
+
+## 0.0.60
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+  - @llamaindex/openai@0.4.4
+
 ## 0.0.59
 
 ### Patch Changes

packages/providers/clip/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/clip",
   "description": "Clip Embedding Adapter for LlamaIndex",
-  "version": "0.0.59",
+  "version": "0.0.61",
   "type": "module",
   "types": "dist/index.d.ts",
   "main": "dist/index.cjs",

packages/providers/cohere/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/cohere
 
+## 0.0.25
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.0.24
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 0.0.23
 
 ### Patch Changes

packages/providers/cohere/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/cohere",
   "description": "Cohere Adapter for LlamaIndex",
-  "version": "0.0.23",
+  "version": "0.0.25",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/deepinfra/CHANGELOG.md

@@ -1,5 +1,24 @@
 # @llamaindex/deepinfra
 
+## 0.0.61
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+  - @llamaindex/openai@0.4.5
+
+## 0.0.60
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+  - @llamaindex/openai@0.4.4
+
 ## 0.0.59
 
 ### Patch Changes

packages/providers/deepinfra/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/deepinfra",
   "description": "Deepinfra Adapter for LlamaIndex",
-  "version": "0.0.59",
+  "version": "0.0.61",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/deepseek/CHANGELOG.md

@@ -1,5 +1,18 @@
 # @llamaindex/deepseek
 
+## 0.0.21
+
+### Patch Changes
+
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/openai@0.4.5
+
+## 0.0.20
+
+### Patch Changes
+
+- @llamaindex/openai@0.4.4
+
 ## 0.0.19
 
 ### Patch Changes

packages/providers/deepseek/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/deepseek",
   "description": "DeepSeek Adapter for LlamaIndex",
-  "version": "0.0.19",
+  "version": "0.0.21",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/discord/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @llamaindex/discord
 
+## 0.1.10
+
+### Patch Changes
+
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.1.9
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
 ## 0.1.8
 
 ### Patch Changes

packages/providers/discord/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/discord",
   "description": "Discord Reader for LlamaIndex",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "type": "module",
   "types": "dist/index.d.ts",
   "main": "dist/index.cjs",

packages/providers/fireworks/CHANGELOG.md

@@ -1,5 +1,18 @@
 # @llamaindex/fireworks
 
+## 0.0.21
+
+### Patch Changes
+
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/openai@0.4.5
+
+## 0.0.20
+
+### Patch Changes
+
+- @llamaindex/openai@0.4.4
+
 ## 0.0.19
 
 ### Patch Changes

packages/providers/fireworks/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/fireworks",
   "description": "Fireworks Adapter for LlamaIndex",
-  "version": "0.0.19",
+  "version": "0.0.21",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/google/CHANGELOG.md

@@ -1,5 +1,29 @@
 # @llamaindex/google
 
+## 0.3.10
+
+### Patch Changes
+
+- d8ac8d3: Feat: add support for openai realtime API
+- Updated dependencies [a89e187]
+- Updated dependencies [62699b7]
+- Updated dependencies [c5b2691]
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/core@0.6.11
+
+## 0.3.9
+
+### Patch Changes
+
+- Updated dependencies [1b5af14]
+  - @llamaindex/core@0.6.10
+
+## 0.3.8
+
+### Patch Changes
+
+- 56763dc: Update to the latest Gemini 2.5 Pro Preview key
+
 ## 0.3.7
 
 ### Patch Changes

packages/providers/google/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/google",
   "description": "Google Adapter for LlamaIndex",
-  "version": "0.3.7",
+  "version": "0.3.10",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",

packages/providers/google/src/live.ts

@@ -12,14 +12,11 @@ import {
   LiveLLM,
   LiveLLMSession,
   type BaseTool,
-  type ChatMessage,
   type LiveConnectConfig,
-  type MessageContentAudioDetail,
-  type MessageContentDetail,
-  type MessageContentImageDataDetail,
-  type MessageContentVideoDetail,
+  type MessageSender,
 } from "@llamaindex/core/llms";
-import { getEnv, uint8ArrayToBase64 } from "@llamaindex/env";
+import { getEnv } from "@llamaindex/env";
+import { GeminiMessageSender } from "./message-sender";
 import { GEMINI_MODEL, type GeminiVoiceName } from "./types";
 import {
   mapBaseToolToGeminiLiveFunctionDeclaration,
@@ -40,6 +37,10 @@ export class GeminiLiveSession extends LiveLLMSession {
     super();
   }
 
+  get messageSender(): MessageSender {
+    return new GeminiMessageSender(this);
+  }
+
   private isInterruptedEvent(event: LiveServerMessage): boolean {
     return event.serverContent?.interrupted === true;
   }
@@ -70,22 +71,6 @@ export class GeminiLiveSession extends LiveLLMSession {
     return event.setupComplete !== undefined;
   }
 
-  private isTextMessage(content: MessageContentDetail) {
-    return content.type === "text";
-  }
-
-  private isAudioMessage(content: MessageContentDetail) {
-    return content.type === "audio";
-  }
-
-  private isImageMessage(content: MessageContentDetail) {
-    return content.type === "image";
-  }
-
-  private isVideoMessage(content: MessageContentDetail) {
-    return content.type === "video";
-  }
-
   //for the tool call event, we need to return the response with function responses
   private async handleToolCallEvent(
     event: LiveServerMessage,
@@ -195,105 +180,11 @@ export class GeminiLiveSession extends LiveLLMSession {
     }
   }
 
-  private sendTextMessage(message: string, role?: string) {
-    this.session?.sendClientContent({
-      turns: [
-        {
-          parts: [{ text: message }],
-          ...(role ? { role } : {}),
-        },
-      ],
-    });
-  }
-
-  private sendAudioMessage(content: MessageContentAudioDetail, role?: string) {
-    if (typeof content.data === "string") {
-      this.session?.sendRealtimeInput({
-        audio: {
-          data: content.data,
-          mimeType: content.mimeType,
-        },
-      });
-    } else {
-      this.session?.sendRealtimeInput({
-        audio: {
-          data: uint8ArrayToBase64(content.data),
-          mimeType: content.mimeType,
-        },
-      });
-    }
-  }
-
-  private sendImageMessage(
-    content: MessageContentImageDataDetail,
-    role?: string,
-  ) {
-    if (typeof content.data === "string") {
-      this.session?.sendRealtimeInput({
-        media: {
-          data: content.data,
-          mimeType: content.mimeType,
-        },
-      });
-    } else {
-      this.session?.sendRealtimeInput({
-        media: {
-          data: uint8ArrayToBase64(content.data),
-          mimeType: content.mimeType,
-        },
-      });
-    }
-  }
-
-  private sendVideoMessage(content: MessageContentVideoDetail, role?: string) {
-    if (typeof content.data === "string") {
-      this.session?.sendRealtimeInput({
-        video: {
-          data: content.data,
-          mimeType: content.mimeType,
-        },
-      });
-    } else {
-      this.session?.sendRealtimeInput({
-        video: {
-          data: uint8ArrayToBase64(content.data),
-          mimeType: content.mimeType,
-        },
-      });
-    }
-  }
-
-  private handleUserInput(message: ChatMessage) {
-    const { content, role } = message;
-
-    if (!Array.isArray(content)) {
-      this.sendTextMessage(content, role);
-    } else {
-      for (const item of content) {
-        if (this.isTextMessage(item)) {
-          this.sendTextMessage(item.text, role);
-        } else if (this.isAudioMessage(item)) {
-          this.sendAudioMessage(item as MessageContentAudioDetail, role);
-        } else if (this.isImageMessage(item)) {
-          this.sendImageMessage(item as MessageContentImageDataDetail, role);
-        } else if (this.isVideoMessage(item)) {
-          this.sendVideoMessage(item as MessageContentVideoDetail, role);
-        }
-      }
-    }
-  }
-
-  sendMessage(message: ChatMessage) {
-    if (!this.session) {
-      throw new Error("Session not connected");
-    }
-    this.handleUserInput(message);
-  }
-
   async disconnect() {
     if (!this.session) {
       throw new Error("Session not connected");
     }
+
     this.session.close();
   }
 }
@@ -322,6 +213,10 @@ export class GeminiLive extends LiveLLM {
     }
   }
 
+  async getEphemeralKey(): Promise<string | undefined> {
+    throw new Error("Ephemeral key is not supported for Gemini Live");
+  }
+
   async connect(config?: LiveConnectConfig) {
     const liveConfig: GoogleLiveConnectConfig = {
       responseModalities: config?.responseModality
@@ -356,6 +251,12 @@ export class GeminiLive extends LiveLLM {
       };
     }
 
+    if (config?.audioConfig) {
+      throw new Error(
+        "Audio config is not supported for Gemini Live, directly send and recieve audio events instead",
+      );
+    }
+
     const geminiLiveSession = new GeminiLiveSession();
 
     geminiLiveSession.session = await this.client.live.connect({

packages/providers/google/src/message-sender.ts

@@ -0,0 +1,53 @@
+import type {
+  MessageContentAudioDetail,
+  MessageContentImageDataDetail,
+  MessageContentVideoDetail,
+  MessageSender,
+} from "@llamaindex/core/llms";
+import type { GeminiLiveSession } from "./live";
+
+export class GeminiMessageSender implements MessageSender {
+  private geminiSession: GeminiLiveSession;
+
+  constructor(session: GeminiLiveSession) {
+    this.geminiSession = session;
+  }
+
+  sendTextMessage(message: string, role?: string) {
+    this.geminiSession.session?.sendClientContent({
+      turns: [
+        {
+          parts: [{ text: message }],
+          ...(role ? { role } : {}),
+        },
+      ],
+    });
+  }
+
+  sendAudioMessage(content: MessageContentAudioDetail, role?: string) {
+    this.geminiSession.session?.sendRealtimeInput({
+      audio: {
+        data: content.data,
+        mimeType: content.mimeType,
+      },
+    });
+  }
+
+  sendImageMessage(content: MessageContentImageDataDetail, role?: string) {
+    this.geminiSession.session?.sendRealtimeInput({
+      media: {
+        data: content.data,
+        mimeType: content.mimeType,
+      },
+    });
+  }
+
+  sendVideoMessage(content: MessageContentVideoDetail, role?: string) {
+    this.geminiSession.session?.sendRealtimeInput({
+      video: {
+        data: content.data,
+        mimeType: content.mimeType,
+      },
+    });
+  }
+}

packages/providers/google/src/types.ts

@@ -75,7 +75,7 @@ export enum GEMINI_MODEL {
   GEMINI_2_0_PRO_EXPERIMENTAL = "gemini-2.0-pro-exp-02-05",
   GEMINI_2_0_FLASH_LIVE = "gemini-2.0-flash-live-001",
   GEMINI_2_5_PRO_PREVIEW = "gemini-2.5-pro-preview-03-25",
-  GEMINI_2_5_PRO_PREVIEW_LATEST = "gemini-2.5-pro-preview-05-06",
+  GEMINI_2_5_PRO_PREVIEW_LATEST = "gemini-2.5-pro-preview-06-05",
   GEMINI_2_5_FLASH_PREVIEW = "gemini-2.5-flash-preview-05-20",
 }
 

packages/providers/groq/CHANGELOG.md

@@ -1,5 +1,18 @@
 # @llamaindex/groq
 
+## 0.0.76
+
+### Patch Changes
+
+- Updated dependencies [d8ac8d3]
+  - @llamaindex/openai@0.4.5
+
+## 0.0.75
+
+### Patch Changes
+
+- @llamaindex/openai@0.4.4
+
 ## 0.0.74
 
 ### Patch Changes

packages/providers/groq/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@llamaindex/groq",
   "description": "Groq Adapter for LlamaIndex",
-  "version": "0.0.74",
+  "version": "0.0.76",
   "type": "module",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",