Now supporting Baidu Wenxin and Baidu AI Search

Ask Baidu’s AI assistant 文心助手 any prompt, or read the AI answer at the top of a Baidu results page, and receive each answer with its sources.

What you get

• BAIDU_WENXIN: the answer of 文心助手, Baidu’s AI assistant at chat.baidu.com. Every result carries text, markdown, sources and searchQueries.
• BAIDU_AI_SEARCH: the AI answer (百度AI+) at the top of a Baidu results page, with card, text, markdown and sources.
• The rest of that results page: organicResults, ads, relatedSearches, contentResults for news, videos and notes, shoppingResults, images, map and otherCards, each when the page shows it.

How it works

Both engines ask Baidu as a first-time visitor in mainland China. 文心助手 searches the web before it answers and lists the pages it read, so sources holds those pages. Asked in English for commuter earbuds, it ran 5 searches, read 26 pages, 19 of them on Sohu and Baijiahao, and answered in English.

Baidu AI Search reads the results page a visitor gets for payload.query. card is general for the 百度AI+ card, where each [1] in markdown matches the position of a source, and health for the 百度健康 card, which adds the review line it shows as reviewedBy. Baidu shows an AI answer on some searches only, so a search with no AI answer after three tries completes with card: null and the rest of the page.

For 降噪耳机哪个牌子好 the AI answer named Sony, Bose and Apple as the top tier and cited 8 sources, Baidu Baike and PCMag among them. The same page carried 6 organic results, 3 ads and 12 related searches.

The example below is a BAIDU_WENXIN task. For Baidu AI Search, set taskType to BAIDU_AI_SEARCH and send the search as payload.query.

Where it helps

• Track how Baidu’s assistant and Baidu search describe your brand to people in China, with prompts in Simplified Chinese.
• Add Baidu to the Chinese prompt sets you already track in Qwen and Doubao.
• Compare the pages Baidu’s AI answer cites with the organic results and ads on the same page.

Send a request

Send the request with your API key in the Authorization header.

curl -X POST https://api.querying.ai/v1/async/task \
  -H "Authorization: Bearer $QUERYING_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "taskType": "BAIDU_WENXIN",
    "payload": {
      "prompt": "Search the web and recommend 3 wireless earbuds for commuting. Compare noise cancellation and battery life in a short table, with source links. Keep the answer under 180 words."
    }
  }'

Request fields

taskTypestringRequired

The engine to run, such as CHATGPT or GOOGLE.

payload.promptstringRequired

The question, written the way a user types it into the chat. Up to 100,000 characters.

webhook.urlstring

An HTTPS address on your server. The finished task is posted there the moment it completes.

idempotencyKeystring

Your own ID for the request. Each key creates one task per account, so a retried send stays a single task.

Response

A finished task looks like this, trimmed for length.

{
  "success": true,
  "task": {
    "id": "aa524d15-1945-491e-bdfc-ef34a0c933e5",
    "taskType": "BAIDU_WENXIN",
    "status": "COMPLETED",
    "createdAt": "2026-11-02T12:51:32.104Z"
  },
  "credits": { "creditsToCharge": 1.25, "creditsCharged": 1.25 },
  "response": {
    "text": "3 Commuting Wireless Earbuds Comparison\n\n| Model | Noise Cancellation | Total Battery Life |\n|-------|--------------------|--------------------|\n| Beats Studio Buds+ | Strong active noise cancellation | 36 hours |\n| Bose QuietComfort Ultra Earbuds | Adaptive smart noise cancellation | 30 hours |\n| EarFun AirPro4+ | AI-enhanced ANC | 48 hours |\n…",
    "markdown": "### 3 Commuting Wireless Earbuds Comparison\n\n| Model | Noise Cancellation | Total Battery Life |\n|-------|--------------------|--------------------|\n| Beats Studio Buds+ | Strong active noise cancellation | 36 hours |\n…",
    "sources": [
      {
        "position": 1,
        "url": "https://www.sohu.com/a/1095420712_122645258",
        "label": "通勤地铁用蓝牙耳机推荐：BOSE、EarFun与SANAG真实体验与场景适配解析_传导_乐图_耳道",
        "source": "搜狐"
      },
      {
        "position": 2,
        "url": "https://www.sohu.com/a/1096142207_122645258",
        "label": "通勤用什么蓝牙耳机好？2026年四大热门机型场景化实测与选购建议_运动_手机_训练",
        "source": "搜狐"
      }
    ],
    "searchQueries": [
      "适合 commuting的无线耳塞推荐",
      "2026年热门通勤无线耳塞推荐",
      "无线耳塞降噪效果和续航对比",
      "…"
    ]
  }
}

task.statusstring

QUEUED, PROCESSING, then COMPLETED with the answer or FAILED with error.

credits.creditsChargedinteger

Credits charged for the finished task. A failed task charges 0.

response.textstring

The answer as plain text.

response.markdownstring

The answer with its headings, lists and tables, in every result.

response.sourcesarray

The pages the answer cites, in the engine’s order, each with position, url and label.

response.searchQueriesarray

The web searches the engine ran to write the answer.

Receive the result

The request returns the task right away with status QUEUED. Poll GET /v1/async/task/{id} until status is COMPLETED or FAILED, or add webhook.url and the finished task arrives at your server with the same task, credits and response. Results stay available for 24 hours after a task finishes.

curl https://api.querying.ai/v1/async/task/$TASK_ID \
  -H "Authorization: Bearer $QUERYING_API_KEY"

Pricing

Each completed task uses 1.25 credits on both engines. A Baidu AI Search task that completes with card: null also uses 1.25 credits, and a failed task uses none.