Box Developer Documentation
Log in

AI agent for structured extract request

The AI agent to be used for structured extraction.

string
ai_agent_extract_structured

The type of AI agent to be used for extraction.

Value is always ai_agent_extract_structured

object

AI agent tool used to handle basic text.

AI LLM endpoint params AWS / AI LLM endpoint params Google / AI LLM endpoint params OpenAI object

The parameters for the LLM endpoint specific to OpenAI / Google models.

string
azure__openai__gpt_3_5_turbo_16k

The model used for the AI agent for basic text. For specific model values, see the available models list.

integer
8400
1

The number of tokens for completion.

string
(\{user_question\}[\s\S]*?\{content\}|\{content\}[\s\S]*?\{user_question\})
It is `{current_date}`, consider these travel options `{content}` and answer the `{user_question}`.
10000

The prompt template contains contextual information of the request and the user prompt. When passing prompt_template parameters, you must include inputs for {user_question} and {content}. {current_date} is optional, depending on the use.

string
You are a helpful travel assistant specialized in budget travel

System messages try to help the LLM "understand" its role and what it is supposed to do.

object

AI agent tool used to to handle longer text.

object
string
openai__text_embedding_ada_002

The model used for the AI agent for calculating embeddings.

object
string
basic

The strategy used for the AI agent for calculating embeddings.

integer
64
1

The number of tokens per chunk.

AI LLM endpoint params AWS / AI LLM endpoint params Google / AI LLM endpoint params OpenAI object

The parameters for the LLM endpoint specific to OpenAI / Google models.

string
azure__openai__gpt_3_5_turbo_16k

The model used for the AI agent for basic text. For specific model values, see the available models list.

integer
8400
1

The number of tokens for completion.

string
(\{user_question\}[\s\S]*?\{content\}|\{content\}[\s\S]*?\{user_question\})
It is `{current_date}`, consider these travel options `{content}` and answer the `{user_question}`.
10000

The prompt template contains contextual information of the request and the user prompt. When passing prompt_template parameters, you must include inputs for {user_question} and {content}. {current_date} is optional, depending on the use.

string
You are a helpful travel assistant specialized in budget travel

System messages try to help the LLM "understand" its role and what it is supposed to do.

Response Example

{
  "type": "ai_agent_extract_structured",
  "basic_text": {
    "llm_endpoint_params": {
      "type": "openai_params",
      "frequency_penalty": 1.5,
      "presence_penalty": 1.5,
      "stop": "<|im_end|>",
      "temperature": 0,
      "top_p": 1
    },
    "model": "azure__openai__gpt_3_5_turbo_16k",
    "num_tokens_for_completion": 8400,
    "prompt_template": "It is `{current_date}`, consider these travel options `{content}` and answer the `{user_question}`.",
    "system_message": "You are a helpful travel assistant specialized in budget travel"
  },
  "long_text": {
    "embeddings": {
      "model": "openai__text_embedding_ada_002",
      "strategy": {
        "id": "basic",
        "num_tokens_per_chunk": 64
      }
    },
    "llm_endpoint_params": {
      "type": "openai_params",
      "frequency_penalty": 1.5,
      "presence_penalty": 1.5,
      "stop": "<|im_end|>",
      "temperature": 0,
      "top_p": 1
    },
    "model": "azure__openai__gpt_3_5_turbo_16k",
    "num_tokens_for_completion": 8400,
    "prompt_template": "It is `{current_date}`, consider these travel options `{content}` and answer the `{user_question}`.",
    "system_message": "You are a helpful travel assistant specialized in budget travel"
  }
}