duck_ai.json

{
  "engine": "duck_ai",
  "params": {
    "q": {
      "required": true,
      "description": "Parameter defines the prompt you want to send to Duck AI, DuckDuckGo's AI chat.",
      "group": "search_query"
    },
    "model": {
      "type": "select",
      "options": [
        "gpt-5.6-luna",
        "gpt-6-luna",
        "gpt-5.4-mini",
        "claude-haiku-4-5",
        "mistral-small-2603",
        "tinfoil/gpt-oss-120b",
        "tinfoil/gemma4-31b"
      ],
      "description": "Parameter defines the AI model that answers the prompt. It can be set to: `gpt-5.6-luna` - GPT-5.6 Luna `gpt-6-luna` - GPT-6 Luna `gpt-5.4-mini` - GPT-5.4 mini `claude-haiku-4-5` - Claude Haiku 4.5 `mistral-small-2603` - Mistral Small 4 `tinfoil/gpt-oss-120b` - gpt-oss 120B `tinfoil/gemma4-31b` - Gemma 4 31B It defaults to `gpt-5.6-luna`.",
      "group": "advanced_parameters"
    },
    "subsequent_request_token": {
      "description": "Parameter defines the token used to continue a Duck AI conversation. Pass this token with a new q to send a follow-up prompt that includes the previous prompts and answers as context (multi-turn). The token is returned by SerpApi from your previous request using our Duck AI API. The token holds the whole conversation, so it grows with each turn. Conversations longer than about 16,000 characters (prompts and answers combined) can't be continued: the request becomes too long and is rejected.",
      "group": "subsequent_request"
    }
  },
  "common_params": {
    "engine": {
      "required": true,
      "description": "Set parameter to `duck_ai` to use the Duck AI API engine.",
      "group": "serpapi_parameters"
    },
    "no_cache": {
      "type": "checkbox",
      "description": "Parameter will force SerpApi to fetch the Duck AI results even if a cached version is already present. A cache is served only if the query and all parameters are exactly the same. Cache expires after 1h. Cached searches are free, and are not counted towards your searches per month. It can be set to `false` (default) to allow results from the cache, or `true` to disallow results from the cache. no\\_cache and async parameters should not be used together.",
      "group": "serpapi_parameters"
    },
    "async": {
      "description": "Parameter defines the way you want to submit your search to SerpApi. It can be set to `false` (default) to open an HTTP connection and keep it open until you got your search results, or `true` to just submit your search to SerpApi and retrieve them later. In this case, you'll need to use our Searches Archive API to retrieve your results. async and no\\_cache parameters should not be used together. async should not be used on accounts with Ludicrous Speed enabled.",
      "group": "serpapi_parameters"
    },
    "zero_trace": {
      "description": "Enterprise only. Parameter enables ZeroTrace mode. It can be set to `false` (default) or `true`. Enable this mode to skip storing search parameters, search files, and search metadata on our servers. This may make debugging more difficult.",
      "group": "serpapi_parameters"
    },
    "api_key": {
      "required": true,
      "description": "Parameter defines the SerpApi private key to use.",
      "group": "serpapi_parameters"
    },
    "output": {
      "description": "Parameter defines the format of the result. It can be set to `json` (default) to get a structured JSON of the results, `html` to get the raw HTML retrieved, or `md` to get a markdown-formatted version optimized for LLMs and AI agents.",
      "group": "serpapi_parameters"
    },
    "json_restrictor": {
      "description": "Parameter defines the fields you want to include in the output, reducing payload size for faster response. See the JSON Restrictor documentation for more details.",
      "group": "serpapi_parameters"
    }
  }
}