Integrating Scaleway Generative APIs with popular AI tools

Scaleway’s Generative APIs are designed to provide easy access to the latest AI models and techniques. Our APIs are built on top of a robust infrastructure that ensures scalability, reliability, and security. With our APIs, you can integrate AI capabilities into your applications, such as text generation, image classification, and more.

Comparison of AI tools and librariesLink to this anchor

The following table compares AI tools and libraries supported by Scaleway’s Generative APIs:

OpenAI client librariesLink to this anchor

Scaleway Generative APIs follow OpenAI’s API structure, making integration straightforward. To get started, you’ll need to install the OpenAI library and set up your API key.

ConfigurationLink to this anchor

To use the OpenAI client library with Scaleway’s Generative APIs, first install the required dependencies:

pip install openai

Then set the API key and base URL in your OpenAI-compatible client:

from openai import OpenAI
client = OpenAI(
    base_url="https://api.scaleway.ai/v1",
    api_key="<API secret key>"
)

Using OpenAI client for text generationLink to this anchor

To use OpenAI client for text generation, you can create a client.chat.completions object and call the create method:

response = client.chat.completions.create(
    model="llama-3.1-8b-instruct",
    messages=[{"role": "user", "content": "Tell me a joke about AI"}]
)
print(response.choices[0].message.content)

LangChain (RAG & LLM applications)Link to this anchor

LangChain is a popular library for building AI applications. Scaleway’s Generative APIs support LangChain for both inference and embeddings.

PythonLink to this anchor

Function calling

1. Run the following commands to install LangChain and its dependencies:

   $ pip install 'langchain>=0.3.24'
   $ pip install 'langchain-core>=0.3.55'
   $ pip install 'langchain-openai>=0.3.14'
   $ pip install 'langchain-text-splitters>=0.3.8'
   

2. Create a file named tools.py and paste the code below into it to import and create the tools examples:

   from langchain_core.messages import HumanMessage
   from langchain.chat_models import init_chat_model
   from langchain_core.tools import tool
   @tool
   def add(a: int, b: int) -> int:
       """Adds a and b."""
       return a + b
   @tool
   def multiply(a: int, b: int) -> int:
       """Multiplies a and b."""
       return a * b
   tools = [add, multiply]
   

3. Configure the init_chat_model function to use Scaleway’s Generative APIs.

   llm = init_chat_model("mistral-small-3.1-24b-instruct-2503", model_provider="openai", base_url="https://api.scaleway.ai/v1")
   

4. Use the llm object and the tools list to generate a response to your query with the following code:

   query = "What is 3 * 12?"
   # You can also try the following query:
   # query = "What is 42 + 4?"
   messages = [HumanMessage(query)]  # We initialize the messages list with the user's query.
   ai_msg = llm_with_tools.invoke(messages)  # We generate a response to the query.
   messages.append(ai_msg)  # We append the response to the messages list.
   for tool_call in ai_msg.tool_calls:
       selected_tool = {"add": add, "multiply": multiply}[tool_call["name"].lower()]  # Depending on the tool name, we select the appropriate tool.
       tool_msg = selected_tool.invoke(tool_call)  # We invoke the selected tool with the tool call.
       messages.append(tool_msg)  # We append the tool's response to the messages list.
   print(llm_with_tools.invoke(messages).content)  # We print the content of the final response.
   

5. Run tools.py:

   $ python tools.py
   The result of 3 * 12 is 36.
   

LlamaIndex (advanced RAG applications)Link to this anchor

LlamaIndex is an open-source framework for building Large Language Models (LLMs) based applications, especially optimizing RAG (Retrieval Augmented Generation) pipelines.

1. Install the required dependencies to use the LlamaIndex framework with Scaleway’s Generative APIs:

   pip install llama-index-llms-openai-like
   

2. Create a file main.py and add the following code to it to configure the OpenAILike client and your secret key:

   from llama_index.llms.openai_like import OpenAILike
   from llama_index.core.llms import ChatMessage
   llm = OpenAILike(
     model="llama-3.1-8b-instruct",
     api_key="<API secret key>",
     api_base="https://api.scaleway.ai/v1",
     max_tokens=512,
     temperature=0.7,
     top_p=1,
     presence_penalty=0,
   )
   

   Tip

   Make sure to replace <API secret key> with your actual API key.

3. You can then interact with the LLM by sending messages to the model with the following code:

   response = llm.chat([ChatMessage("Could you tell me about Scaleway please ?")])
   print(response)
   

4. Finally, run main.py:

   python main.py
   

   The LLM response should display an answer:

   Generally, Scaleway is a reliable and secure cloud provider that offers a range of services for businesses and developers.
   

Javascript (Typescript)Link to this anchor

To perform chat conversations with Langchain, first install langchain and @langchain/openai packages using your node package manager.

1. Use the following command to install Langchain using npm (yarn and pnpm are also available):

   npm install langchain @langchain/openai
   

2. Edit your package.json file to ensure it has the "type": "module" property:

   {
     "type": "module",
     "dependencies": {
       "@langchain/openai": "^0.4.4",
       "langchain": "^0.3.19"
     }
   }
   

3. Create a main.js file and add the following content to it:

   import { ChatOpenAI } from "@langchain/openai";
   const chat = new ChatOpenAI({
       apiKey: "<API secret key>",
       model: "llama-3.1-8b-instruct",
       configuration: {
           baseURL: "https://api.scaleway.ai/v1",
       }
     });
   const response = await chat.invoke("Tell me a joke");
   console.log(response.content);
   

   Tip

   Make sure to replace <API secret key> with your actual API secret key.

4. Run main.js:

   node main.js
   

   The model answer should display:

   Why couldn't the bicycle stand up by itself? Because it was two-tired.
   

Note that other Langchain objects from OpenAI client library are also compatible, such as OpenAI and OpenAIEmbeddings.

Continue Dev (AI coding assistance)Link to this anchor

Continue Dev is a library that provides AI-powered coding assistance. Scaleway’s Generative APIs support Continue Dev for code completion and more.

Zed AI (coding assistance)Link to this anchor

Zed is an IDE (Integrated Development Environment) including AI coding assistance support. Scaleway’s Generative APIs support Zed AI code completion and more.

Chatbox AILink to this anchor

Chatbox AI is a powerful AI client and smart assistant, compatible with Scaleway’s Generative APIs service. It is available across multiple platforms, including Windows, macOS, Android, iOS, Web, and Linux.

Bolt.diy (code generation)Link to this anchor

Bolt.diy is a software enabling users to create web applications from the prompt.

1. Install and launch Bolt.diy locally. Follow the setup instructions provided in theBolt.diy repository.
2. Once Bolt.diy is running, open the interface in your web browser.
3. Click the icon next to the Bolt logo to display the sidebar.
4. Click the cogwheel icon in the sidebar’s bottom left corner to open the Settings.
5. Click Local Providers to add a new external provider configuration.
6. Toggle the switch next to OpenAILike to enable it. Then, enter the Scaleway API endpoint: https://api.scaleway.ai/v1 as the base URL.
7. In Bolt’s main menu, select OpenAILike and input your Scaleway Secret Key as the OpenAILike API Key.
8. Select one of the supported models from Scaleway Generative APIs. For best results with Bolt.diy, which requires a significant amount of output tokens (8000 by default), start with the gemma-3-27b-it model.
9. Enter your prompt in the Bolt.diy interface to see your application being generated.

Alternatively, you can also setup your Scaleway Secret Key by renaming .env.example to .env, adding corresponding environment variables values and restarting Bolt.diy:

OPENAI_LIKE_API_BASE_URL=https://api.scaleway.ai/v1
OPENAI_LIKE_API_KEY=###SCW_SECRET_KEY###

To enable configuration from .env is taken into account, you still need to enable OpenAILike from Bolt interface in Settings > Local Providers section.

Chatbox AILink to this anchor

Chatbox AI is a powerful AI client and smart assistant, compatible with Scaleway’s Generative APIs service. It is available across multiple platforms, including Windows, macOS, Android, iOS, Web, and Linux.

Custom HTTP integrationsLink to this anchor

You can interact with Scaleway’s Generative APIs directly using any HTTP client.

cURL exampleLink to this anchor

To use cURL with Scaleway’s Generative APIs, you can use the following command:

curl https://api.scaleway.ai/v1/chat/completions \
  -H "Authorization: Bearer <API secret key>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama-3.1-8b-instruct",
    "messages": [{"role": "user", "content": "What is quantum computing?"}]
  }'

Python HTTP exampleLink to this anchor

To perform HTTP requests with Scaleway’s Generative APIs, install the requests dependency:

pip install requests

Then, you can use the following code:

import requests
headers = {
    "Authorization": "Bearer <API secret key>",
    "Content-Type": "application/json"
}
data = {
    "model": "llama-3.1-8b-instruct",
    "messages": [{"role": "user", "content": "Explain black holes"}]
}
response = requests.post("https://api.scaleway.ai/v1/chat/completions", json=data, headers=headers)
print(response.json()["choices"][0]["message"]["content"])