Serverless LLM Quickstart

DatabaseMart's Serverless LLM is the go-to inference platform for AI developers seeking a low-cost, reliable, and simple solution for shipping AI models. This is a quickstart guide to help you setup your account and get started.

If you have questions, please check our FAQs first or reach out to our support team online.

How to Create a Dedicated Endpoint

1. Log in to console.databasemart.com

You can log in using Google or GitHub authentication, and a new account will be automatically created on your first login. Alternatively, you can sign up using your email address.

2. Manage API Key

DatabaseMart AI authenticates API access using Bearer authentication with an API Key in the request header, e.g., "Authorization: Bearer {API Key}". If you need to access APIs, you can go to Key Management setting page to create or manage your API keys.

Manage API Key
Create API Key
3. Maintain sufficient credit balance in your account

We offer new users a voucher with some credit to try our products. To add more credit, please visit Billing and Payments and follow the guide on Payment Methods. Also, to avoid any interruption to your service, please ensure your account has sufficient credit balance, and Setup Automatic Top-up is a recommended practice.

Credits Recharge
Credits Recharge 2
Credits Recharge 3
4. Create a Dedicate Endpoint

The Dedicated Endpoint product is designed specifically for AI inference. You only need to choose the container plan and make slight configurations according to your business scenarios to quickly deploy an AI inference service.

As shown below, click the Create Dedicated Endpoint button:

Add dedicated endpoint

Select the appropriate GPU instance based on the large model you want to run and click the Create button:

Choose gpu instance

Enter the Endpoint name, select the model, and click the Confirm button:

Config dedicated endpoint
Select a LLM model
Input endpoint name

Automatically enter the deployment state. Wait for a few minutes and the system will automatically enter the Active state after the deployment is completed.

Endpoint deployment in program
Endpoint deployment completed

Enjoy your Dedicated LLM API

The LLM API interface we provide is OpenAI compatible, so you can choose your favorite method to use your dedicated model. Here are three popular usage methods:
Use curl to Call your LLM API

Copy the code below and remember to modify it to your own URL, API Key, and model information.

export API_KEY="{YOUR DBM AI API Key}"

curl "https://xxxxx.databasemart.ai/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{
    "model": "Qwen/Qwen2.5-0.5B-Instruct",
    "messages": [
        {
            "role": "system",
            "content": "Act like you are a helpful assistant."
        },
        {
            "role": "user",
            "content": "Hi there!"
        }
    ],
    "max_tokens": 512
}'
Use curl to query LLM API
Use Postman to test your LLM API

The third method is to use the Postman API testing tool. Refer to the figure below to configure the Post URL, Content-Type and Authorization Headers, and Request body.

Postaman header config

Then click the Send button to send the request, and the response result will be immediately displayed in the window below.

Postman body config
Postman query LLM API result

The complete interface screenshot is attached below:

Use postman to query LLM API
Use Cherry Studio to Chat with your Dedicated Model

Open the Settings interface, select OpenAI, switch to the On state, and configure the API Key, API Host, and Models.

Cherry studio config

Click the Assistants message icon in the upper left corner, return to the Chat window, select the LLM Model to use, and then you can start AI to ask questions freely.

Cherry studio select a model
Cherry studio chat with a model

Manage your Dedicated Endpoint

Click the Manage button to open the node management page, as shown below.

Manage dedicated endpoint

Check the overview of your dedicated endpoint. If you don't need it anymore, just delete it by clicking the red Delete button.

Dedicated endpoint overview

View the deduction records.

Dedicated endpoint transactions

View the hardware metrics curve of the GPU instance.

Dedicated endpoint hardware metrics

View the frequency and success rate of API calls.

Dedicated endpoint API metrics