How to use Venice API

Venice.ai

2025-02-26

The Venice API allows you to use Venice’s core infrastructure through standard OpenAI API format. The API provides access to cutting-edge AI models for both text and image inference, with a unique cost structure using “VCUs” by staking VVV or paying directly with USD. You can learn more about the Venice API here.

Table of Contents

How to access Venice API for private, uncensored AI inference

Users can access the Venice API in 3 different ways:

Pro Account: Users with a PRO account will gain access to the Venice API within the “Explorer Tier”. This tier has lower rate-limits, and is intended for simple interaction with the API.
VCUs: With Venice’s launch of the VVV token, users who stake tokens within the Venice protocol gain access to a daily AI inference allocation (as well as ongoing staking yield). When staking, users receive VCUs, which represent a portion of the overall Venice compute capacity. You can stake VVV tokens and see your VCU allotment here. Users with positive VCU balance are entitled to “Paid Tier” rate limits.
USD: Users can also opt to deposit USD into their account to pay for API inference the same way that they would on other platforms, like OpenAI or Anthropic. Users with positive USD balance are entitled to “Paid Tier” rate limits.

Venice API cost structure

The API uses standard token-based pricing, denominated in both USD and VCU terms.

These vary by the model selected, and can be found here. Please note that there are currently rate limits in place for API usage, separated by “Explorer Tier” and “Paid Tier”, which are being evaluated on an ongoing basis.

Please reach out to us if you are interested in onboarding, but the rate limits present an issue for you.

How to generate a Venice API Key

Once we get ourselves into the “Explorer” or “Paid” API tier, we’re going to get started by generating our API key.

Head over to the Venice API Dashboard
Scroll down to API Keys and click “Generate New API Key”
Enter the relevant information and click “Generate”, and then save your API Key

Note: For more detailed instructions on API Key generation, go here.

Choosing a model with Venice API

Now that we have our API key, we are going to choose the model we would like to use. Venice has a built-in tool to help facilitate simple requests directly through the website at.

The base URL for listing models is:

‍https://api.venice.ai/api/v1/models

1. Find the section that displays “GET /models” and click “Try it”

‍

2. Paste your API key into the Authorization section, and then choose if you’d like to query for image or text models

3. You will see the box on the top right populate with the associated command that can be used to make the API call. For this example we are using cURL, but you can use Python, JavaScript, PHP, Go or Java from this tool

4. Enter the request into a terminal window, or click “Send” directly within the web page to execute the request

5. You will see the 200 http response with all of the models available through Venice (top image being through the website, bottom image through terminal)

6. Choose the model from the list that you’d like to use, and copy the “id”. This id will be used for selecting your model when you create chat or image prompts

Creating a chat prompt with Venice API

For this section we will send out our first chat prompt to the model. There are various options and settings that can be used within this section. For the purpose of this guide, we will show the simplest example of a simple text prompt

The base URL for text chat is:

https://api.venice.ai/api/v1/chat/completions

1. Go to: https://docs.venice.ai/api-reference/endpoint/chat/completions

2. Find the “POST /chat/completions” section and click “Try it”

3. Enter your API Key that you identified in the earlier section

4. Enter the Model ID that you identified in the earlier section

5. Now we will be adding the “messages”, which provide context to the LLM. The key selections here are the “role”, which is defined as “User”, “Assistant”, “Tool”, and “System. The first system message is typically “You are a helpful assistant.”

To do this, select “System Message - object”, and set the “role” to “system”. Then include the text within “content”

6. Following the system message, you will include the first “user” prompt. You can do this by clicking “Add an item” and then setting the option to “User Message - object”. Select the “role” and “user” and include the user prompt you would like to use within “content”

7. When providing chat context, you will include user prompts, and LLM responses. To do this, click “Add an item” and then set the option to “Assistant Message - object”. Set the “role” as “assistant” and then enter the LLM response within the “content”. We will not use this in our example prompt.

8. When all of your inputs are complete, you will see the associated cURL command generated on the top right. This is the command generated using our settings

curl --request POST \
  --url https://api.venice.ai/api/v1/chat/completions \
  --header 'Authorization: Bearer <your api key> ' \
  --header 'Content-Type: application/json' \
  --data '{
  "model": "llama-3.3-70b",
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful assistant."
    },
    {
      "role": "user",
      "content": "Tell me about AI."
    }
  ]
}'

‍

9. You can choose to click “Send” on the top right corner, or enter this into a terminal window. Once the system executes the command, you will get an http200 response with the following:

{
  "id":"chatcmpl-3fbd0a5b76999f6e65ba7c0c858163ab",
  "object":"chat.completion",
  "created":1739638778,
  "model":"llama-3.3-70b",
  "choices":[
   {
     "index":0,
     "message":{
       "role":"assistant",
       "reasoning_content":null,
       "content":"AI, or Artificial Intelligence, refers to the development of computer systems that can perform tasks that would typically require human intelligence, such as learning, problem-solving, and decision-making. These systems use algorithms and data to make predictions, classify objects, and generate insights. AI has many applications, including image and speech recognition, natural language processing, and expert systems. It can be used in various industries, such as healthcare, finance, and transportation, to improve efficiency and accuracy. AI models can be trained on large datasets to learn patterns and relationships, and they can be fine-tuned to perform specific tasks. Some AI systems, like chatbots and virtual assistants, can interact with humans and provide helpful responses.",
       "tool_calls":[]
   },
   "logprobs":null,
   "finish_reason":"stop",
   "stop_reason":null
  }
],
"usage":{
  "prompt_tokens":483,
  "total_tokens":624,
  "completion_tokens":141,
  "prompt_tokens_details":null
},
  "prompt_logprobs":null
}

‍

10. You just completed your first text prompt using the Venice API!

Creating an image prompt with Venice API

For this section we will send out our first image prompt to the model. There are various image options and settings that can be used in this section, as well as generation or upscaling options. For this example, we will show the simplest example of an image prompt, without styles being selected.

The base URL for image generation is:

https://api.venice.ai/api/v1/image/generate

The base URL for image upscaling is:

https://api.venice.ai/api/v1/image/upscale

1. Go to https://docs.venice.ai/api-reference/endpoint/image/generate

2. Find the “POST /image/generate” section and click “Try it”

3. Enter your API Key that you identified in the earlier section

4. Enter the Model ID that you identified in the earlier section

5. Now we will be adding the “prompt” for the LLM to use to generate the image.

6. There are a variety of other settings that can be configured within this section, we are showing the simplest example. When all of your inputs are complete, you will see the associated cURL command generated on the top right. This is the command generated using out settings

curl --request POST \
  --url https://api.venice.ai/api/v1/image/generate \
  --header 'Authorization: Bearer <your api key> ' \
  --header 'Content-Type: application/json' \
  --data '{
  "model": "fluently-xl",
  "prompt": "Generate an image that best represents AI"
}'

‍

7. You can choose to click “Send” on the top right corner, or enter this into a terminal window. Once the system executes the command, you will get an http200 response with the following:

{
  "request": {
  "width":1024,
  "height":1024,
  "width":30,
  "hide_watermark":false,
  "return_binary":false,
  "seed":-65940141,
  "model":"fluently-xl",
  "prompt":"Generate an image that best represents AI"
},
"images":[ <base64 image data>

‍

Important note: If you prefer to only have the image, rather than the base64 image data, you can change the “return_binary” setting to “true”. If you change this selection, you will only receive the image and not the full JSON response.

8. You just completed your first image prompt using the Venice API!

Start building with Venice API now

There are a ton of settings within the API for both Text and Image generation that will help tailor the response to exactly what you need.

We recommend that advanced users evaluate these settings, and make modifications to optimize your results. Information regarding these settings are available here.

Please reach out to [email protected] or in the #api channel in our Discord with any questions!

‍

Try it Now