An AI agent, capable of utilizing tools and Wrkflows, acts as an autonomous decision-maker by intelligently selecting the necessary processes and inputs to efficiently complete a given task.
To optimize performance, it is recommended to keep the number of tools and Wrkflows defined to a minimum. Models often experience reduced accuracy in selecting the correct tool when the total number exceeds 10–20.
Like all AI content, use with caution; information may be outdated, incomplete, or inaccurate.
Application
OpenAI
Inputs (what you have)
Name | Description | Data Type | Required? | Example |
OpenAI connected account | API credentials for OpenAI | Connected Account | Yes |
|
Model | The model which the AI Agent should use | Predefined Choice List | Yes | GPT 4o Mini |
Prompt | The prompt to generate completions for. Input will be truncated if it exceeds the limit set by the API | Text (Long) | Yes | What is the weather in Montreal? |
System message | The message which sets the context and expectations for a conversation with the AI | Text (Long) | No | You are a helpful weatherperson you job is to get the weather for a given location |
Thread | JSON array of all past messages | Text (Long) | No |
|
Wrk API key | Wrk API key to import existing Wrkflows as tools for the AI Agent | Connected Account | No |
|
Wrkflow UUIDs | Wrkflow UUIDs exposed to the AI Agent. If left blank, all Wrkflows in the account will be exposed | List of Text (Short) | No |
|
Tools | Additional tools to expose to the AI Agent. Configuration details can be found in the OpenAI documentation | Text (Long) | No | [ { "type": "function", "function": { "name": "get_weather", "parameters": { "type": "object", "properties": { "location": { "type": "string" } } } } } ] |
Temperature | Sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic | Number with decimals | No | 0.8 |
Maximum length | The maximum number of tokens to generate in the completion. Defaults to 1024 if not provided | Number | No | 1024 |
Frequency penalty | A number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far | Number with decimals | No | 0.0 |
Presence penalty | A number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far | Number with decimals | No | 0.0 |
Return response as JSON? | If true, the response of the GPT will be set to JSON mode and return valid JSON | True/False | No | False |
Outputs (what you get)
Name | Description | Data Type | Required? | Example |
Thread | All the existing messages in the agent's thread in JSON | Text (Long) | No | Here is the current weather for Montreal ... |
Latest message | The latest message produced by the AI Agent | Text (Long) | No |
|
Tool name | Name of the tool requested by the agent | Text (Short) | No | get_weather |
Tool parameters | JSON object of the parameters required for the tool | Text (Long) | No | {"location":"Montreal"} |
Wrkflow UUID | UUID of the Wrkflow being called within the tool | Text (Short) | No | 7cff6332-bfdd-42f2-8e17-e52c92687fca
|
Wrkflow start values | Wrkflow start values defined by the tool call | Text (Long) | No | {"location":"Montreal"} |
Outcomes
Name | Description |
Success | This status is selected if the job has successfully completed. |
Tool call | This status is selected if agent requires the use of a tool in order or complete the request |
Requirements
N/A