​
🚀 Introduction

This step-by-step guide shows how to call the POST /completions endpoint, interpret the response, and integrate it into your application.

For more code examples, check out the POST /completions endpoint reference.

​
🔧 Prerequisites

Before you start, you will need:

​
📂 1. Making Your First API Call

The POST /completions endpoint generates text based on a prompt. See an example request below:

curl --request POST \
  --url https://api.conteudo.rdstationmentoria.com.br/rest/completions \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "model": "gpt-3.5-turbo",
    "prompt": "Describe the planet Jupiter in 2 sentences.",
    "max_tokens": 60,
    "temperature": 0.5
  }'

If an error occurs, check if your key is correct and if the endpoint is accessible.

​
✅ 2. Example Response

A typical response (HTTP 200) might be:

{
  "completionId": "cpt-abc123",
  "output": {
    "role": "assistant",
    "content": "Jupiter is the largest planet in the solar system, known for its Great Red Spot, a massive storm. It has a thin ring and dozens of moons, including Io and Europa.",
    "toolCall": {}
  },
  "userId": "user-abc123",
  "workspaceId": "wpc-xyz123",
  "createdAt": "2025-07-01T10:05:00.000Z",
  "updatedAt": "2025-07-01T10:05:00.000Z"
}

​
🛠️ 3. Using tool_call with a Custom Function

You can also instruct the model to invoke tools (functions) in a structured way using the toolChoice field and providing a list of tools. This is useful when you want the model to trigger a specific functionality, such as fetching information or executing an external action based on user input.

See an example:

curl --request POST \
  --url https://api.conteudo.rdstationmentoria.com.br/rest/completions \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "messages": [
      {
        "role": "user",
        "content": "I want to know the weather forecast for tomorrow in São Paulo."
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "fetch_weather_forecast",
          "description": "Fetches the current weather forecast for a city",
          "parameters": {
            "type": "object",
            "properties": {
              "city": {
                "type": "string",
                "description": "City name (e.g., São Paulo)"
              },
              "date": {
                "type": "string",
                "description": "Desired date in YYYY-MM-DD format"
              }
            },
            "required": ["city", "date"]
          }
        }
      }
    ],
    "temperature": 0.7,
    "max_tokens": 100
  }'

You can use this approach to connect the model with specific actions, such as fetching external data, calculating values, or executing custom commands.

​
🧾 Example Response with toolCall

When the model decides to use the tool, the response will include a toolCall field with the filled arguments:

{
  "completionId": "cpt-func789",
  "output": {
    "role": "assistant",
    "toolCall": {
      "name": "fetch_weather_forecast",
      "arguments": {
        "city": "São Paulo",
        "date": "2025-07-03"
      }
    }
  }
}

​
🧠 What to Do with This?

Now that you have the extracted arguments, your backend can execute the actual function (fetch_weather_forecast) with the provided parameters. This flow allows the model to act as a natural interface for your internal functionalities.

Test this example by replacing fetch_weather_forecast with specific functions from your domain — such as scheduling a meeting, querying a lead, or fetching real-time data.

​
📘 Next Steps

​
🚀 Introduction

This step-by-step guide shows how to call the POST /completions endpoint, interpret the response, and integrate it into your application.

For more code examples, check out the POST /completions endpoint reference.

​
🔧 Prerequisites

Before you start, you will need:

​
📂 1. Making Your First API Call

The POST /completions endpoint generates text based on a prompt. See an example request below:

curl --request POST \
  --url https://api.conteudo.rdstationmentoria.com.br/rest/completions \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "model": "gpt-3.5-turbo",
    "prompt": "Describe the planet Jupiter in 2 sentences.",
    "max_tokens": 60,
    "temperature": 0.5
  }'

If an error occurs, check if your key is correct and if the endpoint is accessible.

​
✅ 2. Example Response

A typical response (HTTP 200) might be:

{
  "completionId": "cpt-abc123",
  "output": {
    "role": "assistant",
    "content": "Jupiter is the largest planet in the solar system, known for its Great Red Spot, a massive storm. It has a thin ring and dozens of moons, including Io and Europa.",
    "toolCall": {}
  },
  "userId": "user-abc123",
  "workspaceId": "wpc-xyz123",
  "createdAt": "2025-07-01T10:05:00.000Z",
  "updatedAt": "2025-07-01T10:05:00.000Z"
}

​
🛠️ 3. Using tool_call with a Custom Function

You can also instruct the model to invoke tools (functions) in a structured way using the toolChoice field and providing a list of tools. This is useful when you want the model to trigger a specific functionality, such as fetching information or executing an external action based on user input.

See an example:

curl --request POST \
  --url https://api.conteudo.rdstationmentoria.com.br/rest/completions \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "messages": [
      {
        "role": "user",
        "content": "I want to know the weather forecast for tomorrow in São Paulo."
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "fetch_weather_forecast",
          "description": "Fetches the current weather forecast for a city",
          "parameters": {
            "type": "object",
            "properties": {
              "city": {
                "type": "string",
                "description": "City name (e.g., São Paulo)"
              },
              "date": {
                "type": "string",
                "description": "Desired date in YYYY-MM-DD format"
              }
            },
            "required": ["city", "date"]
          }
        }
      }
    ],
    "temperature": 0.7,
    "max_tokens": 100
  }'

You can use this approach to connect the model with specific actions, such as fetching external data, calculating values, or executing custom commands.

​
🧾 Example Response with toolCall

When the model decides to use the tool, the response will include a toolCall field with the filled arguments:

{
  "completionId": "cpt-func789",
  "output": {
    "role": "assistant",
    "toolCall": {
      "name": "fetch_weather_forecast",
      "arguments": {
        "city": "São Paulo",
        "date": "2025-07-03"
      }
    }
  }
}

​
🧠 What to Do with This?

Now that you have the extracted arguments, your backend can execute the actual function (fetch_weather_forecast) with the provided parameters. This flow allows the model to act as a natural interface for your internal functionalities.

Test this example by replacing fetch_weather_forecast with specific functions from your domain — such as scheduling a meeting, querying a lead, or fetching real-time data.

​
📘 Next Steps