Merge 64136b8410 into 701819d018

This reverts commit 68e151b2bd.

README.md

@@ -84,16 +84,18 @@ There are mainly 2 ways of installing Perplexica - With Docker, Without Docker.
 
 3. After cloning, navigate to the directory containing the project files.
 
-4. Rename the `sample.config.toml` file to `config.toml`. For Docker setups, you need only fill in the following fields:
+4. Update environment variables in `docker-compose.yml` file to configure `config.toml`. 
 
-   - `OPENAI`: Your OpenAI API key. **You only need to fill this if you wish to use OpenAI's models**.
+Example: 
-   - `OLLAMA`: Your Ollama API URL. You should enter it as `http://host.docker.internal:PORT_NUMBER`. If you installed Ollama on port 11434, use `http://host.docker.internal:11434`. For other ports, adjust accordingly. **You need to fill this if you wish to use Ollama's models instead of OpenAI's**.
-   - `GROQ`: Your Groq API key. **You only need to fill this if you wish to use Groq's hosted models**.
-   - `ANTHROPIC`: Your Anthropic API key. **You only need to fill this if you wish to use Anthropic models**.
 
-     **Note**: You can change these after starting Perplexica from the settings dialog.
+Below section in `config.toml` can be configured using variables `MODELS_CUSTOM_OPENAI_API_KEY="sk-123456"`, `MODELS_CUSTOM_OPENAI_API_URL="http://localopenai:11134"` and `MODELS_CUSTOM_OPENAI_MODEL_NAME="meta-llama/llama-4"`
 
-   - `SIMILARITY_MEASURE`: The similarity measure to use (This is filled by default; you can leave it as is if you are unsure about it.)
+```toml
+[MODELS.CUSTOM_OPENAI]
+API_KEY = "sk-123456"
+API_URL = "http://localopenai:11134"
+MODEL_NAME = "meta-llama/llama-4"
+```
 
 5. Ensure you are in the directory containing the `docker-compose.yaml` file and execute:
 
@@ -159,6 +161,7 @@ Perplexica runs on Next.js and handles all API requests. It works right away on
 
 [![Deploy to Sealos](https://raw.githubusercontent.com/labring-actions/templates/main/Deploy-on-Sealos.svg)](https://usw.sealos.io/?openapp=system-template%3FtemplateName%3Dperplexica)
 [![Deploy to RepoCloud](https://d16t0pc4846x52.cloudfront.net/deploylobe.svg)](https://repocloud.io/details/?app_id=267)
+[![Run on ClawCloud](https://raw.githubusercontent.com/ClawCloud/Run-Template/refs/heads/main/Run-on-ClawCloud.svg)](https://template.run.claw.cloud/?referralCode=U11MRQ8U9RM4&openapp=system-fastdeploy%3FtemplateName%3Dperplexica)
 
 ## Upcoming Features
 

app.dockerfile

@@ -21,7 +21,9 @@ COPY --from=builder /home/perplexica/.next/static ./public/_next/static
 
 COPY --from=builder /home/perplexica/.next/standalone ./
 COPY --from=builder /home/perplexica/data ./data
+COPY sample.config.toml /home/perplexica/config.toml
+COPY container_entrypoint.sh /home/perplexica/container_entrypoint.sh
 
 RUN mkdir /home/perplexica/uploads
 
-CMD ["node", "server.js"]
+CMD ["bash", "/home/perplexica/container_entrypoint.sh"]

container_entrypoint.sh

@@ -0,0 +1,49 @@
+#!/usr/bin/env bash
+
+CONFIG_TOML_FILE=/home/perplexica/config.toml
+
+TMP_FILE=${CONFIG_TOML_FILE}.tmp
+touch $TMP_FILE
+
+while IFS= read -r line; do
+    # Check if line is a section header (e.g., "[GENERAL]")
+    if [[ "$line" =~ ^\[([^]]+)\] ]]; then
+        current_section="${BASH_REMATCH[1]}"
+        echo "$line" >> "$TMP_FILE"
+        continue
+    fi
+
+    # Skip empty lines and comments
+    if [[ -z "$line" || "$line" =~ ^[[:space:]]*\# ]]; then
+        echo "$line" >> "$TMP_FILE"
+        continue
+    fi
+
+    # Extract key and value (handling quoted values)
+    key=$(echo "$line" | cut -d '=' -f 1 | xargs)
+    value=$(echo "$line" | cut -d '=' -f 2- | xargs)
+
+
+    # Construct the environment variable name in form of SECTION_KEY (e.g., GENERAL_SIMILARITY_MEASURE, MODELS_GEMINI_API_KEY)
+    current_section=$(echo "$current_section" | sed 's/\./_/')
+    env_var_name="${current_section}_${key}"
+
+    # Check if the environment variable exists
+    env_var_value=$(echo "${!env_var_name}")
+    if [ -n "$env_var_value" ]; then
+        new_value="$env_var_value"
+        echo "$key = $new_value" >> "$TMP_FILE"
+    else
+        # Keep original line if no env var exists
+        echo "$line" >> "$TMP_FILE"
+    fi
+
+done < "$CONFIG_TOML_FILE"
+
+# Replace the original file
+mv "$TMP_FILE" "$CONFIG_TOML_FILE"
+
+echo "Config file updated successfully."
+
+# Start server
+node server.js

docker-compose.yaml

@@ -16,6 +16,19 @@ services:
       dockerfile: app.dockerfile
     environment:
       - SEARXNG_API_URL=http://searxng:8080
+      - GENERAL_SIMILARITY_MEASURE="cosine" # "cosine" or "dot"
+      - GENERAL_KEEP_ALIVE="5m"             # How long to keep Ollama models loaded into memory. (Instead of using -1 use "-1m")
+      - MODELS_OPENAI_API_KEY=""
+      - MODELS_GROQ_API_KEY=""
+      - MODELS_ANTHROPIC_API_KEY=""
+      - MODELS_GEMINI_API_KEY=""
+      - MODELS_CUSTOM_OPENAI_API_KEY=""
+      - MODELS_CUSTOM_OPENAI_API_URL=""
+      - MODELS_CUSTOM_OPENAI_MODEL_NAME=""
+      - MODELS_OLLAMA_API_KEY=""            # Ollama API URL - http://host.docker.internal:11434
+      - MODELS_DEEPSEEK_API_KEY=""
+      - MODELS_LM_STUDIO_API_KEY=""         # LM Studio API URL - http://host.docker.internal:1234
+      - API_ENDPOINTS_SEARXNG=""            # SearxNG API URL - http://localhost:32768
     ports:
       - 3000:3000
     networks:
@@ -23,7 +36,6 @@ services:
     volumes:
       - backend-dbstore:/home/perplexica/data
       - uploads:/home/perplexica/uploads
-      - ./config.toml:/home/perplexica/config.toml
     restart: unless-stopped
 
 networks:

package.json

@@ -1,6 +1,6 @@
 {
   "name": "perplexica-frontend",
-  "version": "1.10.1",
+  "version": "1.10.2",
   "license": "MIT",
   "author": "ItzCrazyKns",
   "scripts": {

sample.config.toml

@@ -25,5 +25,8 @@ API_URL = "" # Ollama API URL - http://host.docker.internal:11434
 [MODELS.DEEPSEEK]
 API_KEY = ""
 
+[MODELS.LM_STUDIO]
+API_URL = "" # LM Studio API URL - http://host.docker.internal:1234
+
 [API_ENDPOINTS]
-SEARXNG = "" # SearxNG API URL - http://localhost:32768
+SEARXNG = "" # SearxNG API URL - http://localhost:32768

src/app/api/config/route.ts

@@ -8,6 +8,7 @@ import {
   getOllamaApiEndpoint,
   getOpenaiApiKey,
   getDeepseekApiKey,
+  getLMStudioApiEndpoint,
   updateConfig,
 } from '@/lib/config';
 import {
@@ -51,6 +52,7 @@ export const GET = async (req: Request) => {
 
     config['openaiApiKey'] = getOpenaiApiKey();
     config['ollamaApiUrl'] = getOllamaApiEndpoint();
+    config['lmStudioApiUrl'] = getLMStudioApiEndpoint();
     config['anthropicApiKey'] = getAnthropicApiKey();
     config['groqApiKey'] = getGroqApiKey();
     config['geminiApiKey'] = getGeminiApiKey();
@@ -93,6 +95,9 @@ export const POST = async (req: Request) => {
         DEEPSEEK: {
           API_KEY: config.deepseekApiKey,
         },
+        LM_STUDIO: {
+          API_URL: config.lmStudioApiUrl,
+        },
         CUSTOM_OPENAI: {
           API_URL: config.customOpenaiApiUrl,
           API_KEY: config.customOpenaiApiKey,

src/app/settings/page.tsx

@@ -7,6 +7,7 @@ import { Switch } from '@headlessui/react';
 import ThemeSwitcher from '@/components/theme/Switcher';
 import { ImagesIcon, VideoIcon } from 'lucide-react';
 import Link from 'next/link';
+import { PROVIDER_METADATA } from '@/lib/providers';
 
 interface SettingsType {
   chatModelProviders: {
@@ -20,6 +21,7 @@ interface SettingsType {
   anthropicApiKey: string;
   geminiApiKey: string;
   ollamaApiUrl: string;
+  lmStudioApiUrl: string;
   deepseekApiKey: string;
   customOpenaiApiKey: string;
   customOpenaiApiUrl: string;
@@ -548,8 +550,9 @@ const Page = () => {
                         (provider) => ({
                           value: provider,
                           label:
+                            (PROVIDER_METADATA as any)[provider]?.displayName ||
                             provider.charAt(0).toUpperCase() +
-                            provider.slice(1),
+                              provider.slice(1),
                         }),
                       )}
                     />
@@ -690,8 +693,9 @@ const Page = () => {
                         (provider) => ({
                           value: provider,
                           label:
+                            (PROVIDER_METADATA as any)[provider]?.displayName ||
                             provider.charAt(0).toUpperCase() +
-                            provider.slice(1),
+                              provider.slice(1),
                         }),
                       )}
                     />
@@ -858,6 +862,25 @@ const Page = () => {
                     onSave={(value) => saveConfig('deepseekApiKey', value)}
                   />
                 </div>
+
+                <div className="flex flex-col space-y-1">
+                  <p className="text-black/70 dark:text-white/70 text-sm">
+                    LM Studio API URL
+                  </p>
+                  <Input
+                    type="text"
+                    placeholder="LM Studio API URL"
+                    value={config.lmStudioApiUrl}
+                    isSaving={savingStates['lmStudioApiUrl']}
+                    onChange={(e) => {
+                      setConfig((prev) => ({
+                        ...prev!,
+                        lmStudioApiUrl: e.target.value,
+                      }));
+                    }}
+                    onSave={(value) => saveConfig('lmStudioApiUrl', value)}
+                  />
+                </div>
               </div>
             </SettingsSection>
           </div>

src/components/MessageBox.tsx

@@ -48,6 +48,7 @@ const MessageBox = ({
   const [speechMessage, setSpeechMessage] = useState(message.content);
 
   useEffect(() => {
+    const citationRegex = /\[([^\]]+)\]/g;
     const regex = /\[(\d+)\]/g;
     let processedMessage = message.content;
 
@@ -67,13 +68,36 @@ const MessageBox = ({
     ) {
       setParsedMessage(
         processedMessage.replace(
-          regex,
+          citationRegex,
-          (_, number) =>
+          (_, capturedContent: string) => {
-            `<a href="${
+            const numbers = capturedContent
-              message.sources?.[number - 1]?.metadata?.url
+              .split(',')
-            }" target="_blank" className="bg-light-secondary dark:bg-dark-secondary px-1 rounded ml-1 no-underline text-xs text-black/70 dark:text-white/70 relative">${number}</a>`,
+              .map((numStr) => numStr.trim());
+
+            const linksHtml = numbers
+              .map((numStr) => {
+                const number = parseInt(numStr);
+
+                if (isNaN(number) || number <= 0) {
+                  return `[${numStr}]`;
+                }
+
+                const source = message.sources?.[number - 1];
+                const url = source?.metadata?.url;
+
+                if (url) {
+                  return `<a href="${url}" target="_blank" className="bg-light-secondary dark:bg-dark-secondary px-1 rounded ml-1 no-underline text-xs text-black/70 dark:text-white/70 relative">${numStr}</a>`;
+                } else {
+                  return `[${numStr}]`;
+                }
+              })
+              .join('');
+
+            return linksHtml;
+          },
         ),
       );
+      setSpeechMessage(message.content.replace(regex, ''));
       return;
     }
 

src/lib/config.ts

@@ -1,7 +1,14 @@
-import fs from 'fs';
-import path from 'path';
 import toml from '@iarna/toml';
 
+// Use dynamic imports for Node.js modules to prevent client-side errors
+let fs: any;
+let path: any;
+if (typeof window === 'undefined') {
+  // We're on the server
+  fs = require('fs');
+  path = require('path');
+}
+
 const configFileName = 'config.toml';
 
 interface Config {
@@ -28,6 +35,9 @@ interface Config {
     DEEPSEEK: {
       API_KEY: string;
     };
+    LM_STUDIO: {
+      API_URL: string;
+    };
     CUSTOM_OPENAI: {
       API_URL: string;
       API_KEY: string;
@@ -43,10 +53,17 @@ type RecursivePartial<T> = {
   [P in keyof T]?: RecursivePartial<T[P]>;
 };
 
-const loadConfig = () =>
+const loadConfig = () => {
-  toml.parse(
+  // Server-side only
-    fs.readFileSync(path.join(process.cwd(), `${configFileName}`), 'utf-8'),
+  if (typeof window === 'undefined') {
-  ) as any as Config;
+    return toml.parse(
+      fs.readFileSync(path.join(process.cwd(), `${configFileName}`), 'utf-8'),
+    ) as any as Config;
+  }
+
+  // Client-side fallback - settings will be loaded via API
+  return {} as Config;
+};
 
 export const getSimilarityMeasure = () =>
   loadConfig().GENERAL.SIMILARITY_MEASURE;
@@ -77,6 +94,9 @@ export const getCustomOpenaiApiUrl = () =>
 export const getCustomOpenaiModelName = () =>
   loadConfig().MODELS.CUSTOM_OPENAI.MODEL_NAME;
 
+export const getLMStudioApiEndpoint = () =>
+  loadConfig().MODELS.LM_STUDIO.API_URL;
+
 const mergeConfigs = (current: any, update: any): any => {
   if (update === null || update === undefined) {
     return current;
@@ -109,10 +129,13 @@ const mergeConfigs = (current: any, update: any): any => {
 };
 
 export const updateConfig = (config: RecursivePartial<Config>) => {
-  const currentConfig = loadConfig();
+  // Server-side only
-  const mergedConfig = mergeConfigs(currentConfig, config);
+  if (typeof window === 'undefined') {
-  fs.writeFileSync(
+    const currentConfig = loadConfig();
-    path.join(path.join(process.cwd(), `${configFileName}`)),
+    const mergedConfig = mergeConfigs(currentConfig, config);
-    toml.stringify(mergedConfig),
+    fs.writeFileSync(
-  );
+      path.join(path.join(process.cwd(), `${configFileName}`)),
+      toml.stringify(mergedConfig),
+    );
+  }
 };

src/lib/providers/anthropic.ts

@@ -1,6 +1,11 @@
 import { ChatAnthropic } from '@langchain/anthropic';
 import { ChatModel } from '.';
 import { getAnthropicApiKey } from '../config';
+
+export const PROVIDER_INFO = {
+  key: 'anthropic',
+  displayName: 'Anthropic',
+};
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
 
 const anthropicChatModels: Record<string, string>[] = [

src/lib/providers/deepseek.ts

@@ -3,6 +3,11 @@ import { getDeepseekApiKey } from '../config';
 import { ChatModel } from '.';
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
 
+export const PROVIDER_INFO = {
+  key: 'deepseek',
+  displayName: 'Deepseek AI',
+};
+
 const deepseekChatModels: Record<string, string>[] = [
   {
     displayName: 'Deepseek Chat (Deepseek V3)',

src/lib/providers/gemini.ts

@@ -4,6 +4,11 @@ import {
 } from '@langchain/google-genai';
 import { getGeminiApiKey } from '../config';
 import { ChatModel, EmbeddingModel } from '.';
+
+export const PROVIDER_INFO = {
+  key: 'gemini',
+  displayName: 'Google Gemini',
+};
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
 import { Embeddings } from '@langchain/core/embeddings';
 
@@ -40,8 +45,12 @@ const geminiChatModels: Record<string, string>[] = [
 
 const geminiEmbeddingModels: Record<string, string>[] = [
   {
-    displayName: 'Gemini Embedding',
+    displayName: 'Text Embedding 004',
-    key: 'gemini-embedding-exp',
+    key: 'models/text-embedding-004',
+  },
+  {
+    displayName: 'Embedding 001',
+    key: 'models/embedding-001',
   },
 ];
 

src/lib/providers/groq.ts

@@ -1,6 +1,11 @@
 import { ChatOpenAI } from '@langchain/openai';
 import { getGroqApiKey } from '../config';
 import { ChatModel } from '.';
+
+export const PROVIDER_INFO = {
+  key: 'groq',
+  displayName: 'Groq',
+};
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
 
 const groqChatModels: Record<string, string>[] = [
@@ -72,6 +77,14 @@ const groqChatModels: Record<string, string>[] = [
     displayName: 'Llama 3.2 90B Vision Preview (Preview)',
     key: 'llama-3.2-90b-vision-preview',
   },
+  /* {
+    displayName: 'Llama 4 Maverick 17B 128E Instruct (Preview)',
+    key: 'meta-llama/llama-4-maverick-17b-128e-instruct',
+  }, */
+  {
+    displayName: 'Llama 4 Scout 17B 16E Instruct (Preview)',
+    key: 'meta-llama/llama-4-scout-17b-16e-instruct',
+  },
 ];
 
 export const loadGroqChatModels = async () => {

src/lib/providers/index.ts

@@ -1,18 +1,60 @@
 import { Embeddings } from '@langchain/core/embeddings';
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
-import { loadOpenAIChatModels, loadOpenAIEmbeddingModels } from './openai';
+import {
+  loadOpenAIChatModels,
+  loadOpenAIEmbeddingModels,
+  PROVIDER_INFO as OpenAIInfo,
+  PROVIDER_INFO,
+} from './openai';
 import {
   getCustomOpenaiApiKey,
   getCustomOpenaiApiUrl,
   getCustomOpenaiModelName,
 } from '../config';
 import { ChatOpenAI } from '@langchain/openai';
-import { loadOllamaChatModels, loadOllamaEmbeddingModels } from './ollama';
+import {
-import { loadGroqChatModels } from './groq';
+  loadOllamaChatModels,
-import { loadAnthropicChatModels } from './anthropic';
+  loadOllamaEmbeddingModels,
-import { loadGeminiChatModels, loadGeminiEmbeddingModels } from './gemini';
+  PROVIDER_INFO as OllamaInfo,
-import { loadTransformersEmbeddingsModels } from './transformers';
+} from './ollama';
-import { loadDeepseekChatModels } from './deepseek';
+import { loadGroqChatModels, PROVIDER_INFO as GroqInfo } from './groq';
+import {
+  loadAnthropicChatModels,
+  PROVIDER_INFO as AnthropicInfo,
+} from './anthropic';
+import {
+  loadGeminiChatModels,
+  loadGeminiEmbeddingModels,
+  PROVIDER_INFO as GeminiInfo,
+} from './gemini';
+import {
+  loadTransformersEmbeddingsModels,
+  PROVIDER_INFO as TransformersInfo,
+} from './transformers';
+import {
+  loadDeepseekChatModels,
+  PROVIDER_INFO as DeepseekInfo,
+} from './deepseek';
+import {
+  loadLMStudioChatModels,
+  loadLMStudioEmbeddingsModels,
+  PROVIDER_INFO as LMStudioInfo,
+} from './lmstudio';
+
+export const PROVIDER_METADATA = {
+  openai: OpenAIInfo,
+  ollama: OllamaInfo,
+  groq: GroqInfo,
+  anthropic: AnthropicInfo,
+  gemini: GeminiInfo,
+  transformers: TransformersInfo,
+  deepseek: DeepseekInfo,
+  lmstudio: LMStudioInfo,
+  custom_openai: {
+    key: 'custom_openai',
+    displayName: 'Custom OpenAI',
+  },
+};
 
 export interface ChatModel {
   displayName: string;
@@ -34,6 +76,7 @@ export const chatModelProviders: Record<
   anthropic: loadAnthropicChatModels,
   gemini: loadGeminiChatModels,
   deepseek: loadDeepseekChatModels,
+  lmstudio: loadLMStudioChatModels,
 };
 
 export const embeddingModelProviders: Record<
@@ -44,6 +87,7 @@ export const embeddingModelProviders: Record<
   ollama: loadOllamaEmbeddingModels,
   gemini: loadGeminiEmbeddingModels,
   transformers: loadTransformersEmbeddingsModels,
+  lmstudio: loadLMStudioEmbeddingsModels,
 };
 
 export const getAvailableChatModelProviders = async () => {

src/lib/providers/lmstudio.ts

@@ -0,0 +1,100 @@
+import { getKeepAlive, getLMStudioApiEndpoint } from '../config';
+import axios from 'axios';
+import { ChatModel, EmbeddingModel } from '.';
+
+export const PROVIDER_INFO = {
+  key: 'lmstudio',
+  displayName: 'LM Studio',
+};
+import { ChatOpenAI } from '@langchain/openai';
+import { OpenAIEmbeddings } from '@langchain/openai';
+import { BaseChatModel } from '@langchain/core/language_models/chat_models';
+import { Embeddings } from '@langchain/core/embeddings';
+
+interface LMStudioModel {
+  id: string;
+  name?: string;
+}
+
+const ensureV1Endpoint = (endpoint: string): string =>
+  endpoint.endsWith('/v1') ? endpoint : `${endpoint}/v1`;
+
+const checkServerAvailability = async (endpoint: string): Promise<boolean> => {
+  try {
+    await axios.get(`${ensureV1Endpoint(endpoint)}/models`, {
+      headers: { 'Content-Type': 'application/json' },
+    });
+    return true;
+  } catch {
+    return false;
+  }
+};
+
+export const loadLMStudioChatModels = async () => {
+  const endpoint = getLMStudioApiEndpoint();
+
+  if (!endpoint) return {};
+  if (!(await checkServerAvailability(endpoint))) return {};
+
+  try {
+    const response = await axios.get(`${ensureV1Endpoint(endpoint)}/models`, {
+      headers: { 'Content-Type': 'application/json' },
+    });
+
+    const chatModels: Record<string, ChatModel> = {};
+
+    response.data.data.forEach((model: LMStudioModel) => {
+      chatModels[model.id] = {
+        displayName: model.name || model.id,
+        model: new ChatOpenAI({
+          openAIApiKey: 'lm-studio',
+          configuration: {
+            baseURL: ensureV1Endpoint(endpoint),
+          },
+          modelName: model.id,
+          temperature: 0.7,
+          streaming: true,
+          maxRetries: 3,
+        }) as unknown as BaseChatModel,
+      };
+    });
+
+    return chatModels;
+  } catch (err) {
+    console.error(`Error loading LM Studio models: ${err}`);
+    return {};
+  }
+};
+
+export const loadLMStudioEmbeddingsModels = async () => {
+  const endpoint = getLMStudioApiEndpoint();
+
+  if (!endpoint) return {};
+  if (!(await checkServerAvailability(endpoint))) return {};
+
+  try {
+    const response = await axios.get(`${ensureV1Endpoint(endpoint)}/models`, {
+      headers: { 'Content-Type': 'application/json' },
+    });
+
+    const embeddingsModels: Record<string, EmbeddingModel> = {};
+
+    response.data.data.forEach((model: LMStudioModel) => {
+      embeddingsModels[model.id] = {
+        displayName: model.name || model.id,
+        model: new OpenAIEmbeddings({
+          openAIApiKey: 'lm-studio',
+          configuration: {
+            baseURL: ensureV1Endpoint(endpoint),
+          },
+          modelName: model.id,
+        }) as unknown as Embeddings,
+      };
+    });
+
+    return embeddingsModels;
+  } catch (err) {
+    console.error(`Error loading LM Studio embeddings model: ${err}`);
+    return {};
+  }
+};

src/lib/providers/ollama.ts

@@ -1,6 +1,11 @@
 import axios from 'axios';
 import { getKeepAlive, getOllamaApiEndpoint } from '../config';
 import { ChatModel, EmbeddingModel } from '.';
+
+export const PROVIDER_INFO = {
+  key: 'ollama',
+  displayName: 'Ollama',
+};
 import { ChatOllama } from '@langchain/community/chat_models/ollama';
 import { OllamaEmbeddings } from '@langchain/community/embeddings/ollama';
 

src/lib/providers/openai.ts

@@ -1,6 +1,11 @@
 import { ChatOpenAI, OpenAIEmbeddings } from '@langchain/openai';
 import { getOpenaiApiKey } from '../config';
 import { ChatModel, EmbeddingModel } from '.';
+
+export const PROVIDER_INFO = {
+  key: 'openai',
+  displayName: 'OpenAI',
+};
 import { BaseChatModel } from '@langchain/core/language_models/chat_models';
 import { Embeddings } from '@langchain/core/embeddings';
 
@@ -25,6 +30,18 @@ const openaiChatModels: Record<string, string>[] = [
     displayName: 'GPT-4 omni mini',
     key: 'gpt-4o-mini',
   },
+  {
+    displayName: 'GPT 4.1 nano',
+    key: 'gpt-4.1-nano',
+  },
+  {
+    displayName: 'GPT 4.1 mini',
+    key: 'gpt-4.1-mini',
+  },
+  {
+    displayName: 'GPT 4.1',
+    key: 'gpt-4.1',
+  },
 ];
 
 const openaiEmbeddingModels: Record<string, string>[] = [

src/lib/providers/transformers.ts

@@ -1,5 +1,10 @@
 import { HuggingFaceTransformersEmbeddings } from '../huggingfaceTransformer';
 
+export const PROVIDER_INFO = {
+  key: 'transformers',
+  displayName: 'Hugging Face',
+};
+
 export const loadTransformersEmbeddingsModels = async () => {
   try {
     const embeddingModels = {

src/lib/utils/documents.ts

@@ -64,7 +64,7 @@ export const getDocumentsFromLinks = async ({ links }: { links: string[] }) => {
         const splittedText = await splitter.splitText(parsedText);
         const title = res.data
           .toString('utf8')
-          .match(/<title>(.*?)<\/title>/)?.[1];
+          .match(/<title.*>(.*?)<\/title>/)?.[1];
 
         const linkDocs = splittedText.map((text) => {
           return new Document({