[FEAT] Add native Baidu Search provider for Agent web browsing (#5388)

* feat: add native baidu search provider for agent web browsing

* chore: address baidu search review feedback

* refactor baiduSearch internal util to be locally scoped

---------

Co-authored-by: Timothy Carambat <rambat1010@gmail.com>

docker/.env.example

@@ -378,6 +378,9 @@ GID='1000'
 #------ Bing Search ----------- https://portal.azure.com/
 # AGENT_BING_SEARCH_API_KEY=
 
+#------ Baidu Search ----------- https://cloud.baidu.com/doc/qianfan-api/s/Wmbq4z7e5
+# AGENT_BAIDU_SEARCH_API_KEY=
+
 #------ Serply.io ----------- https://serply.io/
 # AGENT_SERPLY_API_KEY=
 
@@ -448,4 +451,4 @@ GID='1000'
 
 # (optional) Comma-separated list of skills that are auto-approved.
 # This will allow the skill to be invoked without user interaction.
-# AGENT_AUTO_APPROVED_SKILLS=create-pdf-file,create-word-file
+# AGENT_AUTO_APPROVED_SKILLS=create-pdf-file,create-word-file

frontend/src/pages/Admin/Agents/WebSearchSelection/SearchProviderOptions/index.jsx

@@ -248,6 +248,43 @@ export function BingSearchOptions({ settings }) {
   );
 }
 
+export function BaiduSearchOptions({ settings }) {
+  return (
+    <>
+      <p className="text-sm text-white/60 my-2">
+        You can get an API key{" "}
+        <a
+          href="https://cloud.baidu.com/doc/qianfan-api/s/Wmbq4z7e5"
+          target="_blank"
+          rel="noreferrer"
+          className="text-blue-300 underline"
+        >
+          from Baidu AI Cloud Qianfan.
+        </a>
+      </p>
+      <div className="flex gap-x-4">
+        <div className="flex flex-col w-60">
+          <label className="text-white text-sm font-semibold block mb-3">
+            API Key
+          </label>
+          <input
+            type="password"
+            name="env::AgentBaiduSearchApiKey"
+            className="border-none bg-theme-settings-input-bg text-white placeholder:text-theme-settings-input-placeholder text-sm rounded-lg focus:outline-primary-button active:outline-primary-button outline-none block w-full p-2.5"
+            placeholder="Baidu Search API Key"
+            defaultValue={
+              settings?.AgentBaiduSearchApiKey ? "*".repeat(20) : ""
+            }
+            required={true}
+            autoComplete="off"
+            spellCheck={false}
+          />
+        </div>
+      </div>
+    </>
+  );
+}
+
 export function SerplySearchOptions({ settings }) {
   return (
     <>

frontend/src/pages/Admin/Agents/WebSearchSelection/index.jsx

@@ -4,6 +4,7 @@ import SerpApiIcon from "./icons/serpapi.png";
 import SearchApiIcon from "./icons/searchapi.png";
 import SerperDotDevIcon from "./icons/serper.png";
 import BingSearchIcon from "./icons/bing.png";
+import BaiduSearchIcon from "./icons/baidu.png";
 import SerplySearchIcon from "./icons/serply.png";
 import SearXNGSearchIcon from "./icons/searxng.png";
 import TavilySearchIcon from "./icons/tavily.svg";
@@ -24,6 +25,7 @@ import {
   SearchApiOptions,
   SerperDotDevOptions,
   BingSearchOptions,
+  BaiduSearchOptions,
   SerplySearchOptions,
   SearXNGOptions,
   TavilySearchOptions,
@@ -71,6 +73,14 @@ const SEARCH_PROVIDERS = [
     options: (settings) => <BingSearchOptions settings={settings} />,
     description: "Web search powered by the Bing Search API (paid service).",
   },
+  {
+    name: "Baidu Search",
+    value: "baidu-search",
+    logo: BaiduSearchIcon,
+    options: (settings) => <BaiduSearchOptions settings={settings} />,
+    description:
+      "Web search powered by Baidu Search for stronger zh-CN retrieval.",
+  },
   {
     name: "Serply.io",
     value: "serply-engine",

server/.env.example

@@ -382,6 +382,9 @@ TTS_PROVIDER="native"
 #------ Bing Search ----------- https://portal.azure.com/
 # AGENT_BING_SEARCH_API_KEY=
 
+#------ Baidu Search ----------- https://cloud.baidu.com/doc/qianfan-api/s/Wmbq4z7e5
+# AGENT_BAIDU_SEARCH_API_KEY=
+
 #------ Serply.io ----------- https://serply.io/
 # AGENT_SERPLY_API_KEY=
 

server/models/systemSettings.js

@@ -147,6 +147,7 @@ const SystemSettings = {
             "searchapi",
             "serper-dot-dev",
             "bing-search",
+            "baidu-search",
             "serply-engine",
             "searxng-engine",
             "tavily-search",
@@ -483,6 +484,7 @@ const SystemSettings = {
       AgentSearchApiEngine: process.env.AGENT_SEARCHAPI_ENGINE || "google",
       AgentSerperApiKey: !!process.env.AGENT_SERPER_DEV_KEY || null,
       AgentBingSearchApiKey: !!process.env.AGENT_BING_SEARCH_API_KEY || null,
+      AgentBaiduSearchApiKey: !!process.env.AGENT_BAIDU_SEARCH_API_KEY || null,
       AgentSerplyApiKey: !!process.env.AGENT_SERPLY_API_KEY || null,
       AgentSearXNGApiUrl: process.env.AGENT_SEARXNG_API_URL || null,
       AgentTavilyApiKey: !!process.env.AGENT_TAVILY_API_KEY || null,

server/utils/agents/aibitat/plugins/web-browsing.js

@@ -80,6 +80,9 @@ const webBrowsing = {
               case "bing-search":
                 engine = "_bingWebSearch";
                 break;
+              case "baidu-search":
+                engine = "_baiduSearch";
+                break;
               case "serply-engine":
                 engine = "_serplyEngine";
                 break;
@@ -604,6 +607,113 @@ const webBrowsing = {
             );
             return result;
           },
+          _baiduSearch: async function (query) {
+            if (!process.env.AGENT_BAIDU_SEARCH_API_KEY) {
+              this.super.introspect(
+                `${this.caller}: I can't use Baidu Search because the user has not defined the required API key.\nVisit: https://cloud.baidu.com/doc/qianfan-api/s/Wmbq4z7e5 to create the API key.`
+              );
+              return `Search is disabled and no content was found. This functionality is disabled because the user has not set it up yet.`;
+            }
+
+            this.super.introspect(
+              `${this.caller}: Using Baidu Search to search for "${
+                query.length > 100 ? `${query.slice(0, 100)}...` : query
+              }"`
+            );
+
+            const { response, error } = await fetch(
+              "https://qianfan.baidubce.com/v2/ai_search/web_search",
+              {
+                method: "POST",
+                headers: {
+                  "Content-Type": "application/json",
+                  Authorization: `Bearer ${process.env.AGENT_BAIDU_SEARCH_API_KEY}`,
+                  "X-Appbuilder-Authorization": `Bearer ${process.env.AGENT_BAIDU_SEARCH_API_KEY}`,
+                },
+                body: JSON.stringify({
+                  messages: [{ role: "user", content: query }],
+                  resource_type_filter: [{ type: "web", top_k: 10 }],
+                }),
+              }
+            )
+              .then(async (res) => {
+                if (res.ok) return res.json();
+
+                const body = await res.text().catch(() => "");
+                throw new Error(
+                  `${res.status} - ${res.statusText}. params: ${JSON.stringify({
+                    auth: this.middleTruncate(
+                      process.env.AGENT_BAIDU_SEARCH_API_KEY,
+                      5
+                    ),
+                    q: query,
+                    body: body.slice(0, 300),
+                  })}`
+                );
+              })
+              .then((data) => {
+                return { response: data, error: null };
+              })
+              .catch((e) => {
+                this.super.handlerProps.log(`Baidu Search Error: ${e.message}`);
+                return { response: null, error: e.message };
+              });
+
+            if (error)
+              return `There was an error searching for content. ${error}`;
+
+            if (
+              (response?.code || response?.message) &&
+              !response?.references
+            ) {
+              return `There was an error searching for content. ${response?.message || response?.code}`;
+            }
+
+            /**
+             * Normalize Baidu Search References to the expected search results format
+             * @param {Array} references - The references to normalize
+             * @returns {Array} The normalized references
+             */
+            function normalizeBaiduSearchReferences(references = []) {
+              if (!Array.isArray(references)) return [];
+
+              const seenLinks = new Set();
+              return references
+                .filter((reference) => {
+                  if (!reference) return false;
+                  const referenceType = String(
+                    reference.type || reference.resource_type || "web"
+                  ).toLowerCase();
+                  return referenceType === "web";
+                })
+                .map((reference) => {
+                  const title = String(
+                    reference.title || reference.web_anchor || ""
+                  ).trim();
+                  const link = String(reference.url || "").trim();
+                  const snippet = String(
+                    reference.snippet || reference.content || ""
+                  ).trim();
+
+                  if (!title || !link || seenLinks.has(link)) return null;
+                  seenLinks.add(link);
+
+                  return { title, link, snippet };
+                })
+                .filter(Boolean);
+            }
+
+            const data = normalizeBaiduSearchReferences(response?.references);
+            if (data.length === 0)
+              return `No information was found online for the search query.`;
+
+            this.reportSearchResultsCitations(data);
+            const result = JSON.stringify(data);
+            this.super.introspect(
+              `${this.caller}: I found ${data.length} results - reviewing the results now. (~${this.countTokens(result)} tokens)`
+            );
+            return result;
+          },
           _serplyEngine: async function (
             query,
             language = "en",

server/utils/helpers/updateENV.js

@@ -586,6 +586,10 @@ const KEY_MAPPING = {
     envKey: "AGENT_BING_SEARCH_API_KEY",
     checks: [],
   },
+  AgentBaiduSearchApiKey: {
+    envKey: "AGENT_BAIDU_SEARCH_API_KEY",
+    checks: [],
+  },
   AgentSerplyApiKey: {
     envKey: "AGENT_SERPLY_API_KEY",
     checks: [],