Intelligent conversation API

API

The Web Q&A Agent uses web search capabilities to deliver an end-to-end Q&A service, incorporating intent understanding, search, inference, and generation. When you send a user message through the Agent Conversation API, the service performs a sequence of operations, including intent understanding, query rewriting, real-time web retrieval, and tool calling. This process uses an iterative search to generate answers with citations. The service can also return structured cards for weather, stocks, news, sports, and Points of Interest (POIs), along with a recommended follow-up question. Ideal use cases include AI Q&A assistants, in-vehicle conversational interfaces, and mobile AI assistants.

Before using the Agent Conversation API, use the console to create a Web Q&A application, configure its features and policies based on your use case, and obtain the application appId. (Note: Use the console's debugging features to test and improve your application's Q&A performance.)

Request information

The agent chat API provides two response formats:

• Streaming (SSE): stream=true (default). The API sends the reasoning process, search process, incremental content, and structured data in real time as event frames. This format is ideal for conversational interactions.

• Non-streaming (JSON): stream=false. The API returns the complete result in a single response. This format is suitable for batch processing.

Authentication method

1. Activate service and get API key

Log in to the CleverSee console with your Alibaba Cloud account to activate the Web Search service. Currently, RAM users cannot activate the service. You can create or obtain an API key in one of the following ways:

• Method 1: In the CleverSee console, navigate to Web Q&A > API Reference. In the Call with API key section, click Create API key. Note: The API key is displayed only once upon creation. Save it immediately, as it cannot be retrieved after you close the window. For security purposes, creating an API key also creates a new RAM user for your Alibaba Cloud account and grants it permissions for the Web Search service.

• Method 2: With your Alibaba Cloud account, go to the RAM console. Select a RAM user. In the permission management section for that user, add the AliyunCleverSeeAISearchWebSearchAccess permission. Then, on the credential management tab, create an API key. Note: The API key is displayed only once upon creation. Save it immediately, as it cannot be retrieved after you close the window.

2. API authentication

To authenticate with an API key, include the following fields in the request header:

Note: Use date -u to generate the Date header in RFC 1123 format. Set LC_ALL=C to ensure the output uses English abbreviations for weekdays and months. Replace x-acs-apikey with your API key.

Request parameters

  "history": [
    {
      "role": "user",
      "content": "Help me plan a two-day weekend trip to Hangzhou."
    },
    {
      "role": "assistant",
      "content": "I recommend visiting West Lake and Lingyin Temple on the first day, and Xixi Wetland on the second day."
    }
  ]

Note: Multi-turn conversation history is loaded server-side using sessionId. User location is passed via locationInfo.longitude, locationInfo.latitude, and locationInfo.location.

Message

Location info

Option

Important: The following configuration fields are for specific scenarios and must be enabled via allowlist. To use them, contact your Alibaba Cloud sales representative, pre-sales engineer, or after-sales manager. Using these parameters without prior enablement can affect API stability. If these parameters are not enabled or are omitted, the system uses the runtime configuration associated with the appId.

Request example

# Note: The options parameter takes effect only after it has been enabled via allowlist.
# If not enabled, the request ignores the options parameter and uses the configuration of the specified appId.
curl -X POST 'https://cloud-iqs.aliyuncs.com/qa/chat' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -H 'Accept: text/event-stream' \
  -d '{
    "engineType": "LiteAgentAdvanced",
    "appId": "YOUR_APP_ID",
    "sessionId": "example-session-001",
    "message": {
      "role": "user",
      "parts": [
        {
          "type": "text",
          "text": "What is the weather like in Beijing today?"
        }
      ]
    },
    "locationInfo": {
      "longitude": 116.372850,
      "latitude": 39.913828,
      "location": "Xicheng District, Beijing"
    },
    "stream": true,
    "options": {
      "model": "standard",
      "searchMode": "auto",
      "enableRewrite": true,
      "enableCitation": true,
      "enableCards": true,
      "enableFollowUp": false
    }
  }'

Non-streaming response (stream=false)

Non-streaming response parameters

Note: If themessage field is missing from the request body, if no text or image can be parsed frommessage.parts, or if a single Base64 image exceeds 10 MB, the API returns an HTTP 400 error with the structure{ "error": { "message": "..." } }. If the Agent fails to execute, the API returns an HTTP 500 error for the non-streaming response, includingrequestId,messageId,finishReason="error", and anerror object.

1. Thinking mode response comparison

2. Search round

3. Search round item

Successful non-streaming response (model=reasoning_search)

{
  "requestId": "1e2ab45b17857521848613954d4a2e",
  "messageId": "msg-1e2ab45b17857521848613954d4a2e",
  "finishReason": "stop",
  "content": "## Today's Weather in China (Hangzhou)\n\nClear to cloudy, 29°C–40°C. Current temperature is 35°C, feels like 38°C, with a level 4 northeast wind.\n\nLocalized thunderstorms and strong winds may occur in the evening and at night. Take precautions against heatstroke and strong convective weather when outdoors. The air quality is Good, with an AQI of 63.",
  "action": "search",
  "recall": {
    "total": 1,
    "items": [
      {
        "scType": "basic",
        "rankScore": 0.9787976145744324,
        "title": "Weather Forecast for China (Hangzhou) on August 03, 2026",
        "link": "https://tianqi.moji.com/forecast10/china/zhejiang/hangzhou",
        "snippet": "August 03, 2026 China (Hangzhou) weather: Cloudy, temperature 29°/40°.",
        "date": "2026-08-03T14:05:16Z",
        "source": {
          "name": "tianqi.moji.com",
          "domain": "tianqi.moji.com",
          "favicon": ""
        }
      }
    ]
  },
  "searchResults": [
    {
      "sourceType": "web",
      "siteName": "tianqi.moji.com",
      "publishTime": 1785765916,
      "logoUrl": "",
      "id": "52af313215669252",
      "title": "Weather Forecast for China (Hangzhou) on August 03, 2026",
      "snippet": "August 03, 2026 China (Hangzhou) weather: Cloudy, temperature 29°/40°.",
      "url": "https://tianqi.moji.com/forecast10/china/zhejiang/hangzhou",
      "source": {
        "name": "tianqi.moji.com",
        "domain": "tianqi.moji.com"
      },
      "date": "2026-08-03T14:05:16.921+08:00"
    }
  ],
  "cards": [
    {
      "cardType": "weather",
      "cardTitle": "Weather",
      "cardData": {
        "liveWeatherData": {
          "condition": "Clear",
          "conditionId": "5",
          "temperature": 35,
          "realFeel": 38,
          "humidity": 44,
          "icon": "0",
          "uvi": 4,
          "pressure": 1002,
          "visibility": 24.4,
          "wind": {
            "direction": "Northeast",
            "speed": 24.08,
            "level": 4,
            "degrees": 45
          },
          "sunrise": "05:19",
          "sunset": "18:52",
          "tips": "The weather feels hot and dry. Remember to stay hydrated.",
          "updateTime": "2026-08-03 18:10:08"
        },
        "aqiForecast": [
          {
            "date": "2026-08-03",
            "publishTime": "2026-08-03 17:00:00",
            "value": 63,
            "quality": "Good"
          }
        ],
        "city": {
          "name": "Hangzhou",
          "province": "Zhejiang",
          "country": "China",
          "timezone": "Asia/Shanghai"
        },
        "current": {
          "condition": "Clear",
          "conditionId": "5",
          "temperature": 35,
          "realFeel": 38,
          "humidity": 44,
          "icon": "0",
          "uvi": 4,
          "pressure": 1002,
          "visibility": 24.4,
          "wind": {
            "direction": "Northeast",
            "speed": 24.08,
            "level": 4,
            "degrees": 45
          },
          "sunrise": "05:19",
          "sunset": "18:52",
          "tips": "The weather feels hot and dry. Remember to stay hydrated.",
          "updateTime": "2026-08-03 18:10:08"
        },
        "forecast": [
          {
            "date": "2026-08-03",
            "condition": {
              "day": "Cloudy",
              "night": "Clear"
            },
            "iconDay": "1",
            "iconNight": "30",
            "temperature": {
              "day": 40,
              "night": 29
            },
            "humidity": 50,
            "precipitation": {
              "probability": 0,
              "amount": 0
            },
            "uvi": 8,
            "wind": {
              "day": {
                "direction": "Northeast",
                "speed": 15.84,
                "level": 3,
                "degrees": 45
              },
              "night": {
                "direction": "Northeast",
                "speed": 8.64,
                "level": 2,
                "degrees": 45
              }
            },
            "sunrise": "05:19",
            "sunset": "18:52",
            "moonrise": "21:25",
            "moonset": "09:24",
            "moonphase": "WaningGibbous"
          }
        ],
        "aqi": {
          "cityName": "Hangzhou",
          "co": "7",
          "coC": "0.7",
          "no2": "7",
          "no2C": "14.0",
          "o3": "63",
          "o3C": "170.0",
          "pm10": "48",
          "pm10C": "48.0",
          "pm25": "43",
          "pm25C": "30.0",
          "pubtime": "1785747600000",
          "rank": "683/759",
          "so2": "2",
          "so2C": "7.0",
          "value": "63",
          "quality": "Good"
        }
      },
      "index": 0
    }
  ],
  "references": {
    "items": [
      {
        "scType": "basic",
        "sourceType": "web",
        "index": 1,
        "title": "Weather Forecast for China (Hangzhou) on August 03, 2026",
        "url": "https://tianqi.moji.com/forecast10/china/zhejiang/hangzhou",
        "snippet": "August 03, 2026 China (Hangzhou) weather: Cloudy, temperature 29°/40°.",
        "score": 0.9787976145744324,
        "favicon": ""
      }
    ]
  },
  "followUp": {
    "questions": [
      "What is the weather forecast for China (Hangzhou) tomorrow?",
      "When will the yellow alert for thunderstorms and strong winds in China (Hangzhou) be lifted?",
      "Can you provide details on today's Air Quality Index in China (Hangzhou)?"
    ]
  },
  "reasoningContent": "The user is asking for today's weather in China (Hangzhou). Only retain the real-time weather, today's forecast, air quality, and strong convective weather alerts for the current day. Do not provide a multi-day trend.",
  "enableThinking": true,
  "usage": {
    "grandTotalMs": 24768.0,
    "inputTokens": 13198,
    "outputTokens": 1158,
    "totalTokens": 14356
  },
  "code": 200
}

Streaming response (stream=true, SSE)

Event frame format

Business events are streamed in frames using the format data: {JSON}\n\n. The stream is always terminated by a final data: [DONE]\n\n message. While waiting for backend output, the gateway may send an SSE comment frame, such as : heartbeat\n\n. Your client should ignore these comments and continue reading the stream.

Common SSE event fields

SSE event types

1. Streaming sequences for different models

The square brackets [...] indicate a conditional return. By default, the searchMode parameter is set to force, and the action defaults to action=search. The system returns a direct answer (answer) or a clarification (clarify) only when you explicitly pass searchMode=auto.

1.1 model=standard, action=search (no reasoning)

start
→ data-search-results → data-websearch
→ [data-card] × N
→ text-start → text-delta × N → text-end
→ data-reference → [data-follow-up] → data-usage → finish → [DONE]

1.2 model=thinking, action=search (response model enables reasoning)

start
→ data-search-results → data-websearch
→ reasoning-start → reasoning-delta × N → reasoning-end
→ [data-reasoning-reference]
→ [data-card] × N
→ text-start → text-delta × N → text-end
→ data-reference → [data-follow-up] → data-usage → finish → [DONE]

Note: When model=auto_thinking and action=search are specified, the behavior is the same as for thinking if thinking is enabled. Otherwise, the process proceeds directly from data-recall to text-start without sending any reasoning-* or data-reasoning-reference.

1.3 model=reasoning_search, action=search (search-while-reasoning)

start
→ [Round 1: reasoning-start → reasoning-delta × N → reasoning-end
             → data-search-query → data-search-rounds → [data-reasoning-reference]]
→ [Rounds 2...N repeat the events above]
→ data-search-results → data-all-reasoning-reference
→ [data-card] × N
→ text-start → text-delta × N → text-end
→ data-reference → [data-follow-up] → data-usage → finish → [DONE]

1.4 Branches for searchMode=auto

action=answer (standard):
start → text-start → text-delta × N → text-end
      → [data-follow-up] → data-usage → finish → [DONE]


action=answer (thinking / auto_thinking with reasoning activated):
start → reasoning-start → reasoning-delta × N → reasoning-end
      → text-start → text-delta × N → text-end
      → [data-follow-up] → data-usage → finish → [DONE]


action=answer (auto_thinking without reasoning / reasoning_search):
start → text-start → text-delta × N → text-end
      → [data-follow-up] → data-usage → finish → [DONE]


action=clarify (all models):
start → text-start → text-delta × N → text-end
      → data-usage → finish → [DONE]

Note: The action=answer/clarify action does not perform a search. Therefore, it does not generate data-recall, data-reference, or multi-round search events. In this case, the search_rounds field in a non-streaming reasoning_search response is an empty array.

2. Reasoning event examples

data: {"type":"reasoning-start","id":"rsn-req-001-answer","requestId":"req-001"}

data: {"type":"reasoning-delta","id":"rsn-req-001-answer","delta":"First, confirm the time range involved in the question.","requestId":"req-001"}

data: {"type":"reasoning-end","id":"rsn-req-001-answer","requestId":"req-001"}

data: {"type":"data-reasoning-reference","data":{"items":[{"sourceType":"web","sc_type":"basic","index":1,"title":"Source Title","url":"https://example.com/source","snippet":"Source snippet","score":0.95,"favicon":""}]},"requestId":"req-001"}

Note: The only difference between data-reasoning-reference and data-reference is the event type. Their data structures are identical. During search-while-reasoning, it immediately follows the data-search-rounds of the current round, and the frontend assigns it to the current round in event order.

Streaming response frame parameters

1. data-websearch (web search results)

2. data-search-results (web search results + tool call results)

3. Reference event parameters

4. data-follow-up (follow-up recommendation)

5. data-usage (usage)

Example of a successful SSE frame

{
  "event": "text-delta",
  "data": {
    "type": "text-delta",
    "requestId": "2a127bc9-9474-405d-916d-8bc4475fa459",
    "id": "txt-2a127bc9-9474-405d-916d-8bc4475fa459",
    "delta": "It is cloudy in Hangzhou today, with a temperature of 18–25℃."
  },
  "id": "2a127bc9-9474-405d-916d-8bc4475fa459"
}

Card response structure

When a dialog triggers a vertical tool, the API returns a structured card in cards (non-streaming) or in the data-card event (streaming). All cards share the following outer structure:

Top-level structure: Stock query card

{
  "cardType": "stock_quote",
  "cardTitle": "Stock Quote",
  "cardData": {
    "name": "Alibaba",
    "type": "US Stock",
    "latestPrice": 123.456
  }
}

Empty card response

If a non-streaming request returns no cards, the top-level cards field is an empty array:

{

  "cards": [ ]

}
Note: If a streaming request does not generate a card, no data-card event is sent. Setting options.enableCards to false only disables the card display; successful tool results are still included in the reply context or searchResults.

Note: If a streaming request generates no card, no data-card event is sent. Setting options.enableCards=false only disables card rendering; successful tool results are still added to the context and searchResults.

Stock quote (stock_quote)

A stock query card includes the stock name, market type, latest price, previous day's closing price, pricing unit, price change percentage, and historical prices. The card type is always stock_quote. A maximum of 10 historical prices are returned. If a valid latest price and a trading date are available, the system also uses the latest price as the historical price for that date.

Stock query card example

{
  "cardType": "stock_quote",
  "cardTitle": "stock quote",
  "cardData": {
    "sc_type": "stock_quote",
    "name": "Alibaba",
    "type": "US stock",
    "lastestPri": "123.456",
    "yestodEndPri": "121.000",
    "unit": "USD",
    "growthRate": {
      "dodGrowthRate": "2.03%",
      "momGrowthRate": "5.31%",
      "myoyGrowthRate": "18.90%"
    },
    "historyPriList": [
      {
        "date": "2026-07-28",
        "endPri": "120.200"
      },
      {
        "date": "2026-07-29",
        "endPri": "121.000"
      },
      {
        "date": "2026-07-30",
        "endPri": "123.456"
      }
    ]
  }
}

Weather

The weather card includes city information, current weather conditions, a weather forecast, real-time air quality, and an air quality forecast. The weather forecast provides a forecast for up to 7 days, including the current day. The AQI forecast filters out yesterday's data from the upstream source and returns data for the current day and the next 5 days (up to 6 entries).

Weather card

{
  "cardType": "weather",
  "cardTitle": "Weather query",
  "cardData": {
    "city": {
      "name": "Hangzhou",
      "province": "Zhejiang",
      "country": "China",
      "timezone": "Asia/Shanghai"
    },
    "current": {
      "condition": "Light rain",
      "conditionId": "66",
      "temperature": 26,
      "realFeel": 31,
      "humidity": 94,
      "icon": "7",
      "uvi": 0,
      "pressure": 1002,
      "visibility": 14.7,
      "wind": {
        "direction": "Southwest",
        "speed": 4.72,
        "level": 1,
        "degrees": 225
      },
      "sunrise": "05:11",
      "sunset": "19:00",
      "tips": "Slightly warm. Dress accordingly.",
      "updateTime": "2026-07-21 20:00:08"
    },
    "live_weather_data": {
      "condition": "Light rain",
      "conditionId": "66",
      "temperature": 26,
      "realFeel": 31,
      "humidity": 94,
      "icon": "7",
      "uvi": 0,
      "pressure": 1002,
      "visibility": 14.7,
      "wind": {
        "direction": "Southwest",
        "speed": 4.72,
        "level": 1,
        "degrees": 225
      },
      "sunrise": "05:11",
      "sunset": "19:00",
      "tips": "Slightly warm. Dress accordingly.",
      "updateTime": "2026-07-21 20:00:08"
    },
    "forecast": [
      {
        "date": "2026-07-22",
        "condition": {
          "day": "Light rain",
          "night": "Partly cloudy"
        },
        "iconDay": "7",
        "iconNight": "31",
        "temperature": {
          "day": 36,
          "night": 27
        },
        "humidity": 84,
        "precipitation": {
          "probability": 60,
          "amount": 2.4
        },
        "uvi": 4,
        "wind": {
          "day": {
            "direction": "Southwest",
            "speed": 8.64,
            "level": 2,
            "degrees": 225
          },
          "night": {
            "direction": "Southeast",
            "speed": 8.64,
            "level": 2,
            "degrees": 135
          }
        },
        "sunrise": "05:12",
        "sunset": "19:00",
        "moonrise": "13:02",
        "moonset": "23:42",
        "moonphase": "WaxingGibbous"
      }
    ],
    "aqi": {
      "cityName": "Hangzhou",
      "co": "4",
      "coC": "0.4",
      "no2": "18",
      "no2C": "18",
      "o3": "42",
      "o3C": "84",
      "pm10": "22",
      "pm10C": "22",
      "pm25": "18",
      "pm25C": "18",
      "pubtime": "2026-07-22 15:00:00",
      "rank": "23",
      "so2": "5",
      "so2C": "5",
      "value": "42"
    },
    "aqi_forecast": [
      {
        "date": "2026-07-22",
        "publishTime": "2026-07-22 08:00:00",
        "value": "42"
      },
      {
        "date": "2026-07-23",
        "publishTime": "2026-07-22 08:00:00",
        "value": "45"
      },
      {
        "date": "2026-07-24",
        "publishTime": "2026-07-22 08:00:00",
        "value": "48"
      },
      {
        "date": "2026-07-25",
        "publishTime": "2026-07-22 08:00:00",
        "value": "50"
      },
      {
        "date": "2026-07-26",
        "publishTime": "2026-07-22 08:00:00",
        "value": "46"
      },
      {
        "date": "2026-07-27",
        "publishTime": "2026-07-22 08:00:00",
        "value": "44"
      }
    ]
  }
}

News

cardData.newsList contains the trending news items. Each item in cardData.newsList[] has the following structure:

Trending topic card

{
  "cardType": "news",
  "cardTitle": "Trending news",
  "cardData": {
    "total": 2,
    "newsList": [
      {
        "title": "Multiple regions receive new favorable policies",
        "link": "https://news.example.com/article/1",
        "tags": [
          "finance",
          "policy"
        ],
        "snippet": "New government guidelines aim to drive structural adjustments across several industries.",
        "images": "https://img.example.com/news/1.jpg",
        "publishTime": "2026-08-01 09:30:00",
        "showTimeText": "24 hours ago",
        "hostname": "Example Finance News"
      },
      {
        "title": "Tech companies unveil annual product lineup",
        "link": "https://news.example.com/article/2",
        "tags": [
          "technology"
        ],
        "snippet": "",
        "images": "",
        "publishTime": "2026-08-02 08:10:00",
        "showTimeText": "3 hours ago",
        "hostname": "Example Technology Channel"
      }
    ]
  }
}

Sports news (sports_news)

A sports news card includes a title, a details URL, a thumbnail, a comments URL, tags, and a publication time. The card type is always sports_news. A query returns up to 10 valid sports news items, and total specifies the number of news items returned.

A sports news item is filtered out if it has neither a title nor a details link. It is also filtered out if its publication time can be parsed and is 365 days or older than the current server date. No expiration check is performed if the publication time is missing or is unparseable. If all news items are filtered out, the sports news card is not returned.

Sports news card

{
  "cardType": "sports_news",
  "cardTitle": "Sports News",
  "cardData": {
    "total": 2,
    "newsList": [
      {
        "title": "Team Announces New Season Roster",
        "url": "https://sports.example.com/news/1",
        "imageUrl": "https://img.example.com/sports/1.jpg",
        "commentUrl": "https://sports.example.com/news/1/comments",
        "labels": [
          "Basketball",
          "NBA"
        ],
        "date": "2026-07-30 09:30:00"
      },
      {
        "title": "Preseason Schedule Officially Released",
        "url": "https://sports.example.com/news/2",
        "imageUrl": "",
        "commentUrl": "",
        "labels": [
          "Basketball"
        ],
        "date": "2026-07-29 18:20:00"
      }
    ]
  }
}

Nearby search (POI)

Returns detailed information about places such as hotels, restaurants, attractions, and entertainment venues. The card type is always poi, and the POI list is in cardData.places. These places share a common structure, but some fields within category and metadata differ by place type.

Note: Although listed separately in the design document, hotels, restaurants, attractions, and entertainment are consolidated here because they share a consistent field structure. The metadata field is the union of all fields across these categories, while the actual response is trimmed to include only fields with valid data from upstream.

cardData

POI

Nearby search card for a restaurant

{
    "cardType": "poi",
    "cardTitle": "nearby search",
    "cardData": {
        "total": 1,
        "places": [
            {
                "types": "Food & Beverage Services|Chinese Restaurant",
                "metadata": {
                    "score": "4.5",
                    "averageSpend": "49.00",
                    "phone": "1875819****",
                    "dailyOpeningHours": "Open 24 hours",
                    "weeklyOpeningDays": "Mon-Sun 00:00-24:00",
                    "mainTag": "Chinese Cuisine"
                },
                "images": [
                    { "url": "https://store.is.autonavi.com/showpic/18abe252a6b5a77edf6b487f2eec3663" },
                    { "url": "https://aos-comment.amap.com/B0HD9GZGPO/comment/e953f0e337f14f939e12acd03f5e8d68_2048_2048_80.jpg" },
                    { "url": "https://aos-comment.amap.com/B0HD9GZGPO/comment/6740bb8dfc2e3a75c14096d34c176a03_2048_2048_80.jpg" }
                ],
                "districtCode": "330110",
                "address": "Tabian Market Intersection, Qingshan Village, Huanghu Town",
                "districtName": "Yuhang District",
                "provinceCode": "330000",
                "cityCode": "0571",
                "latitude": "30.472464",
                "typeCode": "050100",
                "cityName": "Hangzhou City",
                "name": "Qingqing Restaurant",
                "distanceMeter": "2513",
                "id": "B0HD9GZGPO",
                "provinceName": "Zhejiang Province",
                "longitude": "119.830294",
                "category": "Restaurant"
            }
        ]
    }
},
{
    "cardType": "poi",
    "cardTitle": "nearby search",
    "cardData": {
        "total": 10,
        "places": [
            {
                "types": "Food & Beverage Services|Chinese Restaurant|Hot Pot",
                "metadata": {
                    "score": "4.8",
                    "averageSpend": "120.00",
                    "businessArea": "Xidan",
                    "phone": "010-6617****|010-8094****",
                    "dailyOpeningHours": "10:30-07:00",
                    "weeklyOpeningDays": "Mon-Sun 10:30-07:00",
                    "mainTag": "Hot Pot"
                },
                "images": [
                    { "url": "https://store.is.autonavi.com/showpic/bcb84a02dd0ebb388d01c309af554414?operate=original" },
                    { "url": "https://aos-comment.amap.com/B000A7RYT6/comment/20260316-393b4dba62efda998ef08a9b-4W394uGdptg84NHUtHwFzl.jpg" },
                    { "url": "https://aos-comment.amap.com/B000A7RYT6/comment/DF29504C_A24C_4239_90A5_9740CD915D32_L0_001_1179_77_1767485100289_98892747.jpg" }
                ],
                "districtCode": "110102",
                "address": "7th Floor, 109 Xidan North Street, Jewelry & Wedding Building (Opposite Galeries Lafayette)",
                "districtName": "Xicheng District",
                "provinceCode": "110000",
                "cityCode": "010",
                "latitude": "39.913828",
                "typeCode": "050117",
                "cityName": "Beijing City",
                "name": "Haidilao Hot Pot (Xidan Branch)",
                "distanceMeter": "",
                "id": "B000A7RYT6",
                "provinceName": "Beijing City",
                "longitude": "116.372850",
                "category": "Restaurant"
            },
            {
                "types": "Food & Beverage Services|Chinese Restaurant|Banquet Restaurant",
                "metadata": {
                    "score": "4.6",
                    "businessArea": "Xidan",
                    "phone": "010-6606****|010-6606****",
                    "mainTag": "Comfortable"
                },
                "images": [
                    { "url": "https://store.is.autonavi.com/showpic/81ac50c18b417f51b97eea17c3d90187", "title": "Logo" },
                    { "url": "https://store.is.autonavi.com/showpic/3aff0794a54a8cfcaf721f11f63daf50", "title": "Exterior" },
                    { "url": "https://store.is.autonavi.com/showpic/a5b7d726ec93b29b63eccb3e4bb6f5ce", "title": "Standard Room" }
                ],
                "districtCode": "110102",
                "address": "No. 1, Lane 1, Damucang Hutong (a 420 m walk from Lingjing Hutong Metro Station's Southwest Exit D)",
                "districtName": "Xicheng District",
                "provinceCode": "110000",
                "cityCode": "010",
                "latitude": "39.912908",
                "typeCode": "050101",
                "cityName": "Beijing City",
                "name": "Hanlin Banquet (Beijing Xidan Branch)",
                "distanceMeter": "",
                "id": "B0L0L5A3CD",
                "provinceName": "Beijing City",
                "longitude": "116.371925",
                "category": "Restaurant"
            }
        ]
    }
}

{
    "code": 400,
    "error": {
        "message": "query cannot be empty (message.parts must contain text or file)",
        "errorCode": "INVALID_PARAM"
    }
}

{
  "error": {
    "message": "You are not authorized to perform this action",
    "code": "403"
  }
}