SambaNova Quickstart - SambaNova Documentation

Get started using the SambaNova API in just a few minutes.

Prerequisites

Before you begin, ensure you have:

A SambaCloud account, or access to a SambaStack deployment through your system administrator.
Python 3, Node.js, or curl installed, depending on your preferred integration method.

Get your API key.

To generate an API key, go to the API keys and URLs page. When generating API keys, be sure to save them securely, as they can’t be viewed again.

You can generate and use up to 25 API keys.

Pick a model.

SambaCloud developers can view the available models and details on the SambaCloud model page.SambaStack developers should consult with their system administrator to determine which models are available on their system. Model details can then be viewed on the SambaStack models page.

This guide uses Meta-Llama-3.3-70B-Instruct as an example for the remainder of these steps.

Make an API request.

You can make an inference request in multiple ways. See two examples below:

SambaNova SDK - Use Javascript or Python for a more flexible integration.
OpenAI client library – Use Javascript or Python for a more flexible integration.
CURL command – Send a request directly from the command line.

SambaNova SDK

To get started, choose your preferred programming language. Next, open a terminal and run the command to install the SambaNova SDK.

//ensure you have Node.js installed.
npm install sambanova

# make sure you have Python3 and pip installed
pip install sambanova

Next, copy the following code into a new file.

import SambaNova from "sambanova";

const client = new SambaNova({
  baseURL: "your-sambanova-base-url",
  apiKey: "your-sambanova-api-key",
});

const chatCompletion = await client.chat.completions.create({
  messages: [
    { role: "system", content: "Answer the question in a couple sentences." },
    { role: "user", content: "Share a happy story with me" },
  ],
  model: "Meta-Llama-3.3-70B-Instruct",
});

console.log(chatCompletion.choices[0].message.content);

from sambanova import SambaNova

client = SambaNova(
    base_url="your-sambanova-base-url",
    api_key="your-sambanova-api-key"
)

completion = client.chat.completions.create(
  model="Meta-Llama-3.3-70B-Instruct",
  messages = [
      {"role": "system", "content": "Answer the question in a couple sentences."},
      {"role": "user", "content": "Share a happy story with me"}
    ]
)

print(completion.choices[0].message.content)

After copying the code into the file, replace the placeholder strings "your-sambanova-base-url" and "your-sambanova-api-key" with your actual Base URL and API Key. Then, run the file in a terminal using the command shown below.

node hello-world.js

python hello_world.py

After running the program, you’ll see output similar to the following:

Here’s a happy story: One day, a little girl named Sophie found a lost puppy in her neighborhood and decided to take it home to care for it. As she nursed the puppy back to health, she named it Max and the two became inseparable best friends, going on adventures and playing together every day.

OpenAI client library

To get started, select your preferred programming language. Then, open a terminal window and run the command to install the OpenAI library.

//ensure you have Node.js installed.
npm install openai

# make sure you have Python3 and pip installed
pip install openai

Next, copy the following code into a new file.

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "your-sambanova-base-url",
  apiKey: "your-sambanova-api-key",
});

const chatCompletion = await client.chat.completions.create({
  messages: [
    { role: "system", content: "Answer the question in a couple sentences." },
    { role: "user", content: "Share a happy story with me" },
  ],
  model: "Meta-Llama-3.3-70B-Instruct",
});

console.log(chatCompletion.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="your-sambanova-base-url",
    api_key="your-sambanova-api-key"
)

completion = client.chat.completions.create(
  model="Meta-Llama-3.3-70B-Instruct",
  messages = [
      {"role": "system", "content": "Answer the question in a couple sentences."},
      {"role": "user", "content": "Share a happy story with me"}
    ]
)

print(completion.choices[0].message.content)

Once copied into the file, replace the string fields "your-sambanova-base-url" and "your-sambanova-api-key" with your base URL and API Key values. Then run the file with the command below in a terminal window.

node hello-world.js

python hello_world.py

After you run the program, you’ll see output similar to the following:

Here’s a happy story: One day, a little girl named Sophie found a lost puppy in her neighborhood and decided to take it home to care for it. As she nursed the puppy back to health, she named it Max and the two became inseparable best friends, going on adventures and playing together every day.

CURL command

In a terminal window, run the CURL command to make your first request to the API.

export API_KEY="your-api-key-here"
export URL="your-url-here"

curl -H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"messages": [
{"role": "system", "content": "Answer the question in a couple sentences."},
{"role": "user", "content": "Share a happy story with me"}
],
"stop": ["<|eot_id|>"],
"model": "Meta-Llama-3.3-70B-Instruct",
"stream": true, "stream_options": {"include_usage": true}
}' \
-X POST $URL

Next steps

Now that you can make requests to a model, great potential of building AI-powered applications await. Get inspired of what to build by exploring the AI Starter Kits, a collection of open-source Python projects.

​Prerequisites

​SambaNova SDK

​OpenAI client library

​CURL command

​Next steps

Prerequisites

SambaNova SDK

OpenAI client library

CURL command

Next steps