Revert "Merge branch 'main' of github.com:TypeCellOS/BlockNote"

This reverts commit 86887fa, reversing
changes made to 7d223de.

Author: matthewlipski

CHANGELOG.md

@@ -1,13 +1,3 @@
-## 0.44.0 (2025-12-02)
-
-### 🚀 Features
-
-- **ai:** Abort requests ([#2213](https://github.com/TypeCellOS/BlockNote/pull/2213))
-
-### ❤️ Thank You
-
-- Yousef
-
 ## 0.43.0 (2025-12-01)
 
 ### 🚀 Features

docs/app/(home)/hero/DemoEditor.tsx

@@ -1,24 +1,24 @@
 import {
   BlockNoteSchema,
   combineByGroup,
+  filterSuggestionItems,
   uploadToTmpFilesDotOrg_DEV_ONLY,
 } from "@blocknote/core";
-import { filterSuggestionItems } from "@blocknote/core/extensions";
-import "@blocknote/core/fonts/inter.css";
 import * as locales from "@blocknote/core/locales";
-import { BlockNoteView } from "@blocknote/mantine";
-import "@blocknote/mantine/style.css";
+import "@blocknote/core/fonts/inter.css";
 import {
   getDefaultReactSlashMenuItems,
   SuggestionMenuController,
   useCreateBlockNote,
 } from "@blocknote/react";
+import { BlockNoteView } from "@blocknote/mantine";
 import {
   getMultiColumnSlashMenuItems,
-  multiColumnDropCursor,
   locales as multiColumnLocales,
+  multiColumnDropCursor,
   withMultiColumn,
 } from "@blocknote/xl-multi-column";
+import "@blocknote/mantine/style.css";
 import { useTheme } from "next-themes";
 import { useCallback, useMemo, useState } from "react";
 import YPartyKitProvider from "y-partykit/provider";

docs/content/docs/features/ai/backend-integration.mdx

@@ -11,16 +11,12 @@ The most common (and recommended) setup to integrate BlockNote AI with an LLM is
 ## Default setup (Vercel AI SDK)
 
 The example below closely follows the [basic example from the Vercel AI SDK](https://ai-sdk.dev/docs/ai-sdk-ui/chatbot#example) for Next.js.
-The only difference is that we're retrieving the BlockNote tools from the request body and using the `toolDefinitionsToToolSet` function to convert them to AI SDK tools. We also forward the serialized document state (selection, cursor, block IDs) that BlockNote adds to every user message by calling `injectDocumentStateMessages`. The LLM will now be able to invoke these tools to make modifications to the BlockNote document as requested by the user. The tool calls are forwarded to the client application where they're handled automatically by the AI Extension.
+The only difference is that we're retrieving the BlockNote tools from the request body and using the `toolDefinitionsToToolSet` function to convert them to AI SDK tools. The LLM will now be able to invoke these tools to make modifications to the BlockNote document as requested by the user. The tool calls are forwarded to the client application where they're handled automatically by the AI Extension.
 
 ```ts app/api/chat/route.ts
 import { openai } from "@ai-sdk/openai";
 import { convertToModelMessages, streamText } from "ai";
-import {
-  aiDocumentFormats,
-  injectDocumentStateMessages,
-  toolDefinitionsToToolSet,
-} from "@blocknote/xl-ai/server";
+import { toolDefinitionsToToolSet } from "@blocknote/xl-ai";
 
 // Allow streaming responses up to 30 seconds
 export const maxDuration = 30;
@@ -30,8 +26,7 @@ export async function POST(req: Request) {
 
   const result = streamText({
     model: openai("gpt-4.1"), // see https://ai-sdk.dev/docs/foundations/providers-and-models
-    system: aiDocumentFormats.html.systemPrompt,
-    messages: convertToModelMessages(injectDocumentStateMessages(messages)),
+    messages: convertToModelMessages(messages),
     tools: toolDefinitionsToToolSet(toolDefinitions),
     toolChoice: "required",
   });
@@ -70,7 +65,7 @@ const model = createOpenAICompatible({
 })('model-id');
 
 // ...
- AIExtension({
+ createAIExtension({
     transport: new ClientSideTransport({
         model,
     }),
@@ -108,4 +103,4 @@ You can connect BlockNote AI features with more advanced AI pipelines. You can i
   with BlockNote AI, [get in touch](/about).
 </Callout>
 
-- By default, BlockNote AI sends the entire LLM chat history to the backend. See [the server persistence example](https://github.com/TypeCellOS/BlockNote/tree/main/examples/09-ai/07-server-persistence) for a pattern where the backend stores chat and only the latest message is sent to the backend.
+- By default, BlockNote AI composes the LLM request (messages) based on the user's prompt and passes these to your backend. See [this example](https://github.com/TypeCellOS/BlockNote/blob/main/examples/09-ai/07-server-promptbuilder/src/App.tsx) for an example where composing the LLM request (prompt building) is delegated to the server.

docs/content/docs/features/ai/getting-started.mdx

@@ -25,7 +25,7 @@ import { createBlockNoteEditor } from "@blocknote/core";
 import { BlockNoteAIExtension } from "@blocknote/xl-ai";
 import { en } from "@blocknote/core/locales";
 import { en as aiEn } from "@blocknote/xl-ai/locales";
-import { AIExtension } from "@blocknote/xl-ai";
+import { createAIExtension } from "@blocknote/xl-ai";
 import "@blocknote/xl-ai/style.css"; // add the AI stylesheet
 
 const editor = createBlockNoteEditor({
@@ -34,7 +34,7 @@ const editor = createBlockNoteEditor({
     ai: aiEn, // add default translations for the AI extension
   },
   extensions: [
-    AIExtension({
+    createAIExtension({
       transport: new DefaultChatTransport({
         api: `/api/chat`,
       }),
@@ -44,7 +44,7 @@ const editor = createBlockNoteEditor({
 });
 ```
 
-See the [API Reference](/docs/features/ai/reference) for more information on the `AIExtension` options.
+See the [API Reference](/docs/features/ai/reference) for more information on the `createAIExtension` method.
 
 ## Adding AI UI elements
 
@@ -85,11 +85,7 @@ This example follows the [basic example from the AI SDK](https://ai-sdk.dev/docs
 ```ts app/api/chat/route.ts
 import { openai } from "@ai-sdk/openai";
 import { convertToModelMessages, streamText } from "ai";
-import {
-  aiDocumentFormats,
-  injectDocumentStateMessages,
-  toolDefinitionsToToolSet,
-} from "@blocknote/xl-ai/server";
+import { toolDefinitionsToToolSet } from "@blocknote/xl-ai";
 
 // Allow streaming responses up to 30 seconds
 export const maxDuration = 30;
@@ -99,8 +95,7 @@ export async function POST(req: Request) {
 
   const result = streamText({
     model: openai("gpt-4.1"), // see https://ai-sdk.dev/docs/foundations/providers-and-models
-    system: aiDocumentFormats.html.systemPrompt,
-    messages: convertToModelMessages(injectDocumentStateMessages(messages)),
+    messages: convertToModelMessages(messages),
     tools: toolDefinitionsToToolSet(toolDefinitions),
     toolChoice: "required",
   });
@@ -109,12 +104,6 @@ export async function POST(req: Request) {
 }
 ```
 
-This follows the regular `streamText` pattern of the AI SDK, with 3 exceptions:
-
-- the BlockNote document state is extracted from message metadata and injected into the messages, using `injectDocumentStateMessages`
-- BlockNote client-side tool definitions are extracted from the request body and passed to the LLM using `toolDefinitionsToToolSet`
-- The system prompt is set to the default BlockNote system prompt (`aiDocumentFormats.html.systemPrompt`). You can override or extend the system prompt. If you do so, make sure your modified system prompt still explains the AI on how to modify the BlockNote document.
-
 See [Backend integrations](/docs/features/ai/backend-integration) for more information on how to integrate BlockNote AI with your backend.
 
 # Full Example