Quickstart

Follow this guide to create an API key and send your first chat request with gpt-5.6-sol. Before you begin, make sure that you can access the Vergora console and that your available balance is sufficient for a test request.

Before you begin

You also need a terminal that can run cURL. The commands below work on macOS, Linux, and environments that support Bash or Zsh.

1. Sign in and open the console

1. Select Log in in the website header.
2. Complete account verification as instructed on the page.
3. Open the console and confirm that you are using the correct account and workspace.
4. Check your available balance and the status of any existing API keys.

If a verification code is incorrect or does not arrive, check the email address, spam folder, and page message before requesting another code.

2. Create an API key

1. Open API Keys in the console.
2. Create a new API key.
3. Use a name that identifies its application and environment, such as demo-development.
4. Follow the page instructions to copy and securely store the complete API key.

An API key is a private credential. Do not expose it in browser code, public repositories, chat messages, logs, or screenshots.

3. Store the API key in your terminal

Use the following command to enter the API key. The terminal does not display the key while you type.

printf "Vergora API Key: "
read -r -s VERGORA_API_KEY
printf "\n"
export VERGORA_API_KEY

This environment variable is available only in the current terminal session. You need to set it again after closing the terminal. Production applications should store credentials in server-side environment variables or a secret manager.

4. Send your first request

Run the following command in the same terminal:

curl --silent --show-error --include \
  "https://api.vergora.ai/v1/chat/completions" \
  -H "Authorization: Bearer ${VERGORA_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [
      {
        "role": "user",
        "content": "Introduce yourself in one sentence."
      }
    ],
    "stream": false
  }'

This example uses the Chat Completions endpoint. Before switching models, confirm that the target model supports the same endpoint, message structure, and parameters.

5. Read the response

A non-streaming Chat Completions-compatible response typically includes message content and usage information. The following fields and values only illustrate where to read the result:

{
  "id": "example-response-id",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Hello! I am an AI assistant."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 8,
    "total_tokens": 20
  }
}

Check the following:

1. The HTTP request completed successfully.
2. The response body contains model output.
3. Your application can read choices[0].message.content.
4. The response includes usage information or an identifier that can help locate the request.

Do not treat the absence of a terminal error as proof that the request succeeded.

6. Review the request record

Open Requests in the console. Find the request by its time and model, then review:

If the record is not visible yet, refresh the page and check the active filters before sending the same request again.

Troubleshooting

After completing your first request, continue to Authentication to learn how to use and protect API keys.