Quickstart

Revision 1 · Updated 2026-11-12 · 2 min read

Send your first model request using the console's connection instructions.

Prerequisites#

Sign in to the Runtoken console. Make sure your account can use the selected model and has available funds.

Have a terminal or application environment ready. The console provides cURL, Python and Node.js connection code.

Create a platform API key#

Open API Keys and choose Create API key.

1. Name the key

   Enter a name that identifies where the key will be used.

2. Set its scope and limits

   Check the expiry, model scope and spend limits before creating the key.

3. Save the key

   Choose Create and show key and save the full key in your secret store.

The full key is shown only when created. For available restrictions and key handling, see Authentication & keys.

Choose a model and connect#

Open Model Market and select a model. Check Supported interfaces in its details.

Open Quick start and select Chat Completions for a model that supports it. Choose cURL, Python or Node.js, then use Copy code.

Before running the code, check these connection details.

Prepare the required runtime and dependencies for the selected language. Run the code in your local or server environment.

Check the response and request record#

Confirm that the response contains model output. Then open Logs.

Set the time range to include the request. Filter by model and API key to find its record.

Choose View to inspect the request ID, protocol, usage and charged amount. Success is the execution status. Settled is the billing status.

If no record appears, widen Time range. Choose 30d to see the past 30 days. Use Refresh logs and check the model and key filters.

If the record is still missing, confirm that the request used the address and key from the same console environment.

Troubleshooting#

Read the returned error before changing the configuration.

Use the integration guide for protocol details. Use OpenAI Chat for conversation history and streamed output.