Fix: terminate onnx inference session manually (#10076)

### What problem does this PR solve?

terminate onnx inference session and release memory manually.

Issue #5050 
Issue #9992 
Issue #8805

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

conf/llm_factories.json

@@ -219,6 +219,70 @@
                 }
             ]
         },
+        {
+            "name": "TokenPony",
+            "logo": "",
+            "tags": "LLM",
+            "status": "1",
+            "llm": [
+                {
+                    "llm_name": "qwen3-8b",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-v3-0324",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "qwen3-32b",
+                    "tags": "LLM,CHAT,131k",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "kimi-k2-instruct",
+                    "tags": "LLM,CHAT,128K",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-r1-0528",
+                    "tags": "LLM,CHAT,164k",
+                    "max_tokens": 164000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "qwen3-coder-480b",
+                    "tags": "LLM,CHAT,1024k",
+                    "max_tokens": 1024000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "glm-4.5",
+                    "tags": "LLM,CHAT,131K",
+                    "max_tokens": 131000,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "deepseek-v3.1",
+                    "tags": "LLM,CHAT,128k",
+                    "max_tokens": 128000,
+                    "model_type": "chat",
+                    "is_tools": true
+                }
+            ]
+        },
         {
             "name": "Tongyi-Qianwen",
             "logo": "",

deepdoc/vision/ocr.py

@@ -13,7 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-
+import gc
 import logging
 import copy
 import time
@@ -348,6 +348,12 @@ class TextRecognizer:
 
         return img
 
+    def close(self):
+        # close session and release manually
+        logging.info('Close TextRecognizer.')
+        del self.predictor
+        gc.collect()
+
     def __call__(self, img_list):
         img_num = len(img_list)
         # Calculate the aspect ratio of all text bars
@@ -395,6 +401,9 @@ class TextRecognizer:
 
         return rec_res, time.time() - st
 
+    def __del__(self):
+        self.close()
+
 
 class TextDetector:
     def __init__(self, model_dir, device_id: int | None = None):
@@ -479,6 +488,11 @@ class TextDetector:
         dt_boxes = np.array(dt_boxes_new)
         return dt_boxes
 
+    def close(self):
+        logging.info("Close TextDetector.")
+        del self.predictor
+        gc.collect()
+
     def __call__(self, img):
         ori_im = img.copy()
         data = {'image': img}
@@ -508,6 +522,9 @@ class TextDetector:
 
         return dt_boxes, time.time() - st
 
+    def __del__(self):
+        self.close()
+
 
 class OCR:
     def __init__(self, model_dir=None):

deepdoc/vision/recognizer.py

@@ -13,7 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-
+import gc
 import logging
 import os
 import math
@@ -406,6 +406,11 @@ class Recognizer:
             "score": float(scores[i])
         } for i in indices]
 
+    def close(self):
+        logging.info("Close recognizer.")
+        del self.ort_sess
+        gc.collect()
+
     def __call__(self, image_list, thr=0.7, batch_size=16):
         res = []
         images = []
@@ -430,5 +435,7 @@ class Recognizer:
 
         return res
 
+    def __del__(self):
+        self.close()
 
 

docs/guides/agent/agent_component_reference/agent.mdx

@@ -26,6 +26,84 @@ An **Agent** component is essential when you need the LLM to assist with summari
 
 2. If your Agent involves dataset retrieval, ensure you [have properly configured your target knowledge base(s)](../../dataset/configure_knowledge_base.md).
 
+## Quickstart
+
+### 1. Click on an **Agent** component to show its configuration panel  
+
+The corresponding configuration panel appears to the right of the canvas. Use this panel to define and fine-tune the **Agent** component's behavior.
+
+### 2. Select your model
+
+Click **Model**, and select a chat model from the dropdown menu. 
+
+:::tip NOTE
+If no model appears, check if your have added a chat model on the **Model providers** page.
+:::
+
+### 3. Update system prompt (Optional)
+
+The system prompt typically defines your model's role. You can either keep the system prompt as is or customize it to override the default.
+
+
+### 4. Update user prompt
+
+The user prompt typically defines your model's task. You will find the `sys.query` variable auto-populated. Type `/` or click **(x)** to view or add variables.
+
+In this quickstart, we assume your **Agent** component is used standalone (without tools or sub-Agents below), then you may also need to specify retrieved chunks using the `formalized_content` variable:
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/standalone_user_prompt_variable.jpg)
+
+### 5. Skip Tools and Agent
+
+The **+ Add tools** and **+ Add agent** sections are used *only* when you need to configure your **Agent** component as a planner (with tools or sub-Agents beneath). In this quickstart, we assume your **Agent** component is used standalone (without tools or sub-Agents beneath). 
+
+### 6. Choose the next component
+
+When necessary, click the **+** button on the **Agent** component to choose the next component in the worflow from the dropdown list.
+
+## Connect to an MCP server as a client
+
+:::danger IMPORTANT
+In this section, we assume your **Agent** will be configured as a planner, with a Tavily tool beneath it.
+:::
+
+### 1. Navigate to the MCP configuration page
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/mcp_page.jpg)
+
+### 2. Configure your Tavily MCP server 
+
+Update your MCP server's name, URL (including the API key), server type, and other necessary settings. When configured correctly, the available tools will be displayed.
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/edit_mcp_server.jpg)
+
+### 3. Navigate to your Agent's editing page
+
+### 4. Connect to your MCP server
+
+1. Click **+ Add tools**:
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/add_tools.jpg)
+
+2. Click **MCP** to show the available MCP servers.
+
+3. Select your MCP server:
+
+   *The target MCP server appears below your Agent component, and your Agent will autonomously decide when to invoke the available tools it offers.*
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/choose_tavily_mcp_server.jpg)
+
+### 5. Update system prompt to specify trigger conditions (Optional)
+
+To ensure reliable tool calls, you may specify within the system prompt which tasks should trigger each tool call.
+
+### 6. View the availabe tools of your MCP server
+
+On the canvas, click the newly-populated Tavily server to view and select its available tools:
+
+![](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/tavily_mcp_server.jpg)
+
+
 ## Configurations
 
 ### Model
@@ -69,7 +147,7 @@ An **Agent** component relies on keys (variables) to specify its data inputs. It
 
 #### Advanced usage
 
-From v0.20.5 onwards, four framework-level prompt blocks are available in the **System prompt** field. Type `/` or click **(x)** to view them; they appear under the **Framework** entry in the dropdown menu.
+From v0.20.5 onwards, four framework-level prompt blocks are available in the **System prompt** field, enabling you to customize and *override* prompts at the framework level. Type `/` or click **(x)** to view them; they appear under the **Framework** entry in the dropdown menu.
 
 - `task_analysis` prompt block
   - This block is responsible for analyzing tasks — either a user task or a task assigned by the lead Agent when the **Agent** component is acting as a Sub-Agent.
@@ -100,6 +178,12 @@ From v0.20.5 onwards, four framework-level prompt blocks are available in the **
 - `citation_guidelines` prompt block
   - Reference design: [citation_prompt.md](https://github.com/infiniflow/ragflow/blob/main/rag/prompts/citation_prompt.md)
 
+*The screenshots below show the framework prompt blocks available to an **Agent** component, both as a standalone and as a planner (with a Tavily tool below):*
+
+![standalone](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/standalone_agent_framework_block.jpg)
+
+![planner](https://raw.githubusercontent.com/infiniflow/ragflow-docs/main/images/planner_agent_framework_blocks.jpg)
+
 ### User prompt
 
 The user-defined prompt. Defaults to `sys.query`, the user query. As a general rule, when using the **Agent** component as a standalone module (not as a planner), you usually need to specify the corresponding **Retrieval** component’s output variable (`formalized_content`) here as part of the input to the LLM.
@@ -129,7 +213,7 @@ Defines the maximum number of attempts the agent will make to retry a failed tas
 
 The waiting period in seconds that the agent observes before retrying a failed task, helping to prevent immediate repeated attempts and allowing system conditions to improve. Defaults to 1 second.
 
-### Max rounds
+### Max reflection rounds
 
 Defines the maximum number reflection rounds of the selected chat model. Defaults to 1 round.
 

docs/references/python_api_reference.md

@@ -977,7 +977,7 @@ The languages that should be translated into, in order to achieve keywords retri
 
 ##### metadata_condition: `dict`
 
-filter condition for meta_fields
+filter condition for `meta_fields`.
 
 #### Returns
 

docs/release_notes.md

@@ -28,11 +28,11 @@ Released on September 10, 2025.
 
 ### Improvements
 
-- Agent Performance Optimized: Improved planning and reflection speed for simple tasks; optimized concurrent tool calls for parallelizable scenarios, significantly reducing overall response time.
+- Agent: 
-- Agent Prompt Framework exposed: Developers can now customize and override framework-level prompts in the system prompt section, enhancing flexibility and control.
+  - Agent Performance Optimized: Improves planning and reflection speed for simple tasks; optimizes concurrent tool calls for parallelizable scenarios, significantly reducing overall response time.
-- Execute SQL Component Enhanced: Replaced the original variable reference component with a text input field, allowing free-form SQL writing with variable support.
+  - Four framework-level prompt blocks are available in the **System prompt** section, enabling customization and overriding of prompts at the framework level, thereby enhancing flexibility and control. See [here](./guides/agent/agent_component_reference/agent.mdx#advanced-usage).
-- Chat: Re-enabled Reasoning and Cross-language search.
+  - **Execute SQL** component enhanced: Replaces the original variable reference component with a text input field, allowing users to write free-form SQL queries and reference variables.
-- Retrieval API Enhanced: Added metadata filtering support to the [Retrieve chunks](https://ragflow.io/docs/dev/http_api_reference#retrieve-chunks) method.
+- Chat: Re-enables **Reasoning** and **Cross-language search**.
 
 ### Added models
 
@@ -45,7 +45,21 @@ Released on September 10, 2025.
 
 - Dataset: Deleted files remained searchable.
 - Chat: Unable to chat with an Ollama model.
-- Agent: Resolved issues including cite toggle failure, task mode requiring dialogue triggers, repeated answers in multi-turn dialogues, and duplicate summarization of parallel execution results.
+- Agent:
+  - A **Cite** toggle failure.
+  - An Agent in task mode still required a dialogue to trigger.
+  - Repeated answers in multi-turn dialogues.
+  - Duplicate summarization of parallel execution results.
+
+### API changes
+
+#### HTTP APIs
+
+- Adds a body parameter `"metadata_condition"` to the [Retrieve chunks](./references/http_api_reference.md#retrieve-chunks) method, enabling metadata-based chunk filtering during retrieval. [#9877](https://github.com/infiniflow/ragflow/pull/9877)
+
+#### Python APIs
+
+- Adds a parameter `metadata_condition` to the [Retrieve chunks](./references/python_api_reference.md#retrieve-chunks) method, enabling metadata-based chunk filtering during retrieval. [#9877](https://github.com/infiniflow/ragflow/pull/9877)
 
 ## v0.20.4
 

rag/flow/parser/parser.py

@@ -45,7 +45,10 @@ class ParserParam(ProcessParamBase):
             "ppt": [],
             "image": [],
             "email": [],
-            "text": [],
+            "text": [
+                "text",
+                "json"
+            ],
             "audio": [],
             "video": [],
         }
@@ -84,7 +87,12 @@ class ParserParam(ProcessParamBase):
                 "parse_method": "ocr",
             },
             "email": {},
-            "text": {},
+            "text": {
+                "suffix": [
+                    "txt"
+                ],
+                "output_format": "json",
+            },
             "audio": {},
             "video": {},
         }
@@ -119,6 +127,11 @@ class ParserParam(ProcessParamBase):
             image_parse_method = image_config.get("parse_method", "")
             self.check_valid_value(image_parse_method.lower(), "Parse method abnormal.", ["ocr"])
 
+        text_config = self.setups.get("text", "")
+        if text_config:
+            text_output_format = text_config.get("output_format", "")
+            self.check_valid_value(text_output_format, "Text output format abnormal.", self.allowed_output_format["text"])
+
     def get_input_form(self) -> dict[str, dict]:
         return {}
 
@@ -208,15 +221,13 @@ class Parser(ProcessBase):
         from rag.app.naive import Markdown as naive_markdown_parser
         from rag.nlp import concat_img
 
-        self.callback(random.randint(1, 5) / 100.0, "Start to work on a Word Processor Document")
+        self.callback(random.randint(1, 5) / 100.0, "Start to work on a markdown.")
 
         blob = from_upstream.blob
         name = from_upstream.name
         conf = self._param.setups["markdown"]
         self.set_output("output_format", conf["output_format"])
 
-        print("markdown {conf=}", flush=True)
-
         markdown_parser = naive_markdown_parser()
         sections, tables = markdown_parser(name, blob, separate_tables=False)
 
@@ -240,13 +251,33 @@ class Parser(ProcessBase):
 
             self.set_output("json", json_results)
 
+    def _text(self, from_upstream: ParserFromUpstream):
+        from deepdoc.parser.utils import get_text
+
+        self.callback(random.randint(1, 5) / 100.0, "Start to work on a text.")
+
+        blob = from_upstream.blob
+        name = from_upstream.name
+        conf = self._param.setups["text"]
+        self.set_output("output_format", conf["output_format"])
+
+        # parse binary to text
+        text_content = get_text(name, binary=blob)
+
+        if conf.get("output_format") == "json":
+            result = [{"text": text_content}]
+            self.set_output("json", result)
+        else:
+            result = text_content
+            self.set_output("text", result)
 
     async def _invoke(self, **kwargs):
         function_map = {
             "pdf": self._pdf,
             "markdown": self._markdown,
             "spreadsheet": self._spreadsheet,
-            "word": self._word
+            "word": self._word,
+            "text": self._text,
         }
         try:
             from_upstream = ParserFromUpstream.model_validate(kwargs)

rag/flow/tests/dsl_examples/general_pdf_all.json

@@ -44,9 +44,12 @@
                     "markdown"
                   ],
                   "output_format": "json"
+                },
+                "text": {
+                  "suffix": ["txt"],
+                  "output_format": "json"
                 }
               }
-            }
           }
         },
         "downstream": ["Chunker:0"],

rag/llm/chat_model.py

@@ -1356,6 +1356,14 @@ class Ai302Chat(Base):
         super().__init__(key, model_name, base_url, **kwargs)
 
 
+class TokenPonyChat(Base):
+    _FACTORY_NAME = "TokenPony"
+
+    def __init__(self, key, model_name, base_url="https://ragflow.vip-api.tokenpony.cn/v1", **kwargs):
+        if not base_url:
+            base_url = "https://ragflow.vip-api.tokenpony.cn/v1"
+
+            
 class MeituanChat(Base):
     _FACTORY_NAME = "Meituan"
 

rag/llm/embedding_model.py

@@ -751,7 +751,11 @@ class SILICONFLOWEmbed(Base):
         token_count = 0
         for i in range(0, len(texts), batch_size):
             texts_batch = texts[i : i + batch_size]
-            texts_batch = [" " if not text.strip() else text for text in texts_batch]
+            if self.model_name in ["BAAI/bge-large-zh-v1.5", "BAAI/bge-large-en-v1.5"]:
+                # limit 512, 340 is almost safe
+                texts_batch = [" " if not text.strip() else truncate(text, 340) for text in texts_batch]
+            else:
+                texts_batch = [" " if not text.strip() else text for text in texts_batch]
 
             payload = {
                 "model": self.model_name,

web/src/components/embed-dialog/index.tsx

@@ -139,7 +139,7 @@ function EmbedDialog({
             </form>
           </Form>
           <div>
-            <span>Embed code</span>
+            <span>{t('embedCode', { keyPrefix: 'search' })}</span>
             <HightLightMarkdown>{text}</HightLightMarkdown>
           </div>
           <div className=" font-medium mt-4 mb-1">

web/src/constants/llm.ts

@@ -54,6 +54,7 @@ export enum LLMFactory {
   DeepInfra = 'DeepInfra',
   Grok = 'Grok',
   XAI = 'xAI',
+  TokenPony = 'TokenPony',
   Meituan = 'Meituan',
 }
 
@@ -114,5 +115,6 @@ export const IconMap = {
   [LLMFactory.DeepInfra]: 'deepinfra',
   [LLMFactory.Grok]: 'grok',
   [LLMFactory.XAI]: 'xai',
+  [LLMFactory.TokenPony]: 'token-pony',
   [LLMFactory.Meituan]: 'longcat',
 };

web/src/locales/zh.ts

@@ -632,6 +632,8 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       },
       cancel: '取消',
       chatSetting: '聊天设置',
+      avatarHidden: '隐藏头像',
+      locale: '地区',
     },
     setting: {
       profile: '概要',

web/src/pages/agent/chat/box.tsx

@@ -62,7 +62,7 @@ function AgentChatBox() {
 
   return (
     <>
-      <section className="flex flex-1 flex-col px-5 h-[90vh]">
+      <section className="flex flex-1 flex-col px-5 min-h-0 pb-4">
         <div className="flex-1 overflow-auto" ref={messageContainerRef}>
           <div>
             {/* <Spin spinning={sendLoading}> */}

web/src/pages/agent/chat/chat-sheet.tsx

@@ -9,7 +9,7 @@ export function ChatSheet({ hideModal }: IModalProps<any>) {
   return (
     <Sheet open modal={false} onOpenChange={hideModal}>
       <SheetContent
-        className={cn('top-20 p-0')}
+        className={cn('top-20 bottom-0 p-0 flex flex-col h-auto')}
         onInteractOutside={(e) => e.preventDefault()}
       >
         <SheetTitle className="hidden"></SheetTitle>

web/src/pages/agent/form/agent-form/index.tsx

@@ -145,7 +145,7 @@ function AgentForm({ node }: INextOperatorForm) {
                   <PromptEditor
                     {...field}
                     placeholder={t('flow.messagePlaceholder')}
-                    showToolbar={false}
+                    showToolbar={true}
                     extraOptions={extraOptions}
                   ></PromptEditor>
                 </FormControl>
@@ -166,7 +166,7 @@ function AgentForm({ node }: INextOperatorForm) {
                     <section>
                       <PromptEditor
                         {...field}
-                        showToolbar={false}
+                        showToolbar={true}
                       ></PromptEditor>
                     </section>
                   </FormControl>

web/src/pages/dataset/sidebar/index.tsx

@@ -9,13 +9,7 @@ import { cn, formatBytes } from '@/lib/utils';
 import { Routes } from '@/routes';
 import { formatPureDate } from '@/utils/date';
 import { isEmpty } from 'lodash';
-import {
+import { Banknote, Database, FileSearch2, GitGraph } from 'lucide-react';
-  Banknote,
-  Database,
-  DatabaseZap,
-  FileSearch2,
-  GitGraph,
-} from 'lucide-react';
 import { useMemo } from 'react';
 import { useTranslation } from 'react-i18next';
 import { useHandleMenuClick } from './hooks';
@@ -34,11 +28,11 @@ export function SideBar({ refreshCount }: PropType) {
 
   const items = useMemo(() => {
     const list = [
-      {
+      // {
-        icon: DatabaseZap,
+      //   icon: DatabaseZap,
-        label: t(`knowledgeDetails.overview`),
+      //   label: t(`knowledgeDetails.overview`),
-        key: Routes.DataSetOverview,
+      //   key: Routes.DataSetOverview,
-      },
+      // },
       {
         icon: Database,
         label: t(`knowledgeDetails.dataset`),

web/src/pages/datasets/dataset-creating-dialog.tsx

@@ -17,16 +17,9 @@ import {
 import { Input } from '@/components/ui/input';
 import { IModalProps } from '@/interfaces/common';
 import { zodResolver } from '@hookform/resolvers/zod';
-import { useForm, useWatch } from 'react-hook-form';
+import { useForm } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
 import { z } from 'zod';
-import {
-  DataExtractKnowledgeItem,
-  DataFlowItem,
-  EmbeddingModelItem,
-  ParseTypeItem,
-  TeamItem,
-} from '../dataset/dataset-setting/configuration/common-item';
 
 const FormId = 'dataset-creating-form';
 
@@ -54,10 +47,6 @@ export function InputForm({ onOk }: IModalProps<any>) {
   function onSubmit(data: z.infer<typeof FormSchema>) {
     onOk?.(data.name);
   }
-  const parseType = useWatch({
-    control: form.control,
-    name: 'parseType',
-  });
   return (
     <Form {...form}>
       <form
@@ -84,15 +73,6 @@ export function InputForm({ onOk }: IModalProps<any>) {
             </FormItem>
           )}
         />
-        <EmbeddingModelItem line={2} />
-        <ParseTypeItem />
-        {parseType === 2 && (
-          <>
-            <DataFlowItem />
-            <DataExtractKnowledgeItem />
-            <TeamItem />
-          </>
-        )}
       </form>
     </Form>
   );

web/src/pages/datasets/dataset-dataflow-creating-dialog.tsx

@@ -0,0 +1,123 @@
+import { ButtonLoading } from '@/components/ui/button';
+import {
+  Dialog,
+  DialogContent,
+  DialogFooter,
+  DialogHeader,
+  DialogTitle,
+} from '@/components/ui/dialog';
+import {
+  Form,
+  FormControl,
+  FormField,
+  FormItem,
+  FormLabel,
+  FormMessage,
+} from '@/components/ui/form';
+import { Input } from '@/components/ui/input';
+import { IModalProps } from '@/interfaces/common';
+import { zodResolver } from '@hookform/resolvers/zod';
+import { useForm, useWatch } from 'react-hook-form';
+import { useTranslation } from 'react-i18next';
+import { z } from 'zod';
+import {
+  DataExtractKnowledgeItem,
+  DataFlowItem,
+  EmbeddingModelItem,
+  ParseTypeItem,
+  TeamItem,
+} from '../dataset/dataset-setting/configuration/common-item';
+
+const FormId = 'dataset-creating-form';
+
+export function InputForm({ onOk }: IModalProps<any>) {
+  const { t } = useTranslation();
+
+  const FormSchema = z.object({
+    name: z
+      .string()
+      .min(1, {
+        message: t('knowledgeList.namePlaceholder'),
+      })
+      .trim(),
+    parseType: z.number().optional(),
+  });
+
+  const form = useForm<z.infer<typeof FormSchema>>({
+    resolver: zodResolver(FormSchema),
+    defaultValues: {
+      name: '',
+      parseType: 1,
+    },
+  });
+
+  function onSubmit(data: z.infer<typeof FormSchema>) {
+    onOk?.(data.name);
+  }
+  const parseType = useWatch({
+    control: form.control,
+    name: 'parseType',
+  });
+  return (
+    <Form {...form}>
+      <form
+        onSubmit={form.handleSubmit(onSubmit)}
+        className="space-y-6"
+        id={FormId}
+      >
+        <FormField
+          control={form.control}
+          name="name"
+          render={({ field }) => (
+            <FormItem>
+              <FormLabel>
+                <span className="text-destructive mr-1"> *</span>
+                {t('knowledgeList.name')}
+              </FormLabel>
+              <FormControl>
+                <Input
+                  placeholder={t('knowledgeList.namePlaceholder')}
+                  {...field}
+                />
+              </FormControl>
+              <FormMessage />
+            </FormItem>
+          )}
+        />
+        <EmbeddingModelItem line={2} />
+        <ParseTypeItem />
+        {parseType === 2 && (
+          <>
+            <DataFlowItem />
+            <DataExtractKnowledgeItem />
+            <TeamItem />
+          </>
+        )}
+      </form>
+    </Form>
+  );
+}
+
+export function DatasetCreatingDialog({
+  hideModal,
+  onOk,
+  loading,
+}: IModalProps<any>) {
+  const { t } = useTranslation();
+
+  return (
+    <Dialog open onOpenChange={hideModal}>
+      <DialogContent className="sm:max-w-[425px]">
+        <DialogHeader>
+          <DialogTitle>{t('knowledgeList.createKnowledgeBase')}</DialogTitle>
+        </DialogHeader>
+        <InputForm onOk={onOk}></InputForm>
+        <DialogFooter>
+          <ButtonLoading type="submit" form={FormId} loading={loading}>
+            {t('common.save')}
+          </ButtonLoading>
+        </DialogFooter>
+      </DialogContent>
+    </Dialog>
+  );
+}

web/src/pages/user-setting/setting-model/ollama-modal/index.tsx

@@ -37,6 +37,7 @@ const llmFactoryToUrlMap = {
     'https://huggingface.co/docs/text-embeddings-inference/quick_tour',
   [LLMFactory.GPUStack]: 'https://docs.gpustack.ai/latest/quickstart',
   [LLMFactory.VLLM]: 'https://docs.vllm.ai/en/latest/',
+  [LLMFactory.TokenPony]: 'https://docs.tokenpony.cn/#/',
 };
 type LlmFactory = keyof typeof llmFactoryToUrlMap;