Mastering Mistral Agents API: Build a Smart Math Solver Agent

Unlocking the Power of Mistral Agents API

The Mistral Agents API empowers developers to craft intelligent, modular agents capable of handling diverse tasks. It supports multimodal models for text and image interactions, conversation memory to maintain context, and flexible engagement with individual or multiple agents.

Key features include built-in tools such as code execution, web browsing, image generation, and document libraries. A unique agent handoff feature allows seamless task passing between agents.

Creating Your First Math-Solving Agent

This guide walks you through building a math-solving agent using Mistral’s API, leveraging the code interpreter tool to solve problems programmatically.

Step 1: Install Dependencies

Begin by installing the Mistral library:

pip install mistralai

Load your API key securely:

from getpass import getpass
apiKey = getpass('Enter Mistral API Key: ')

Step 2: Initialize Client and Create Agent

Set up the Mistral client and define a custom agent named "Math Helper" that solves math problems, evaluates expressions, and explains concepts. It uses the "mistral-medium-2505" model and the built-in code interpreter tool.

from mistralai import Mistral
client = Mistral(apiKey)
math_agent = client.beta.agents.create(
    model="mistral-medium-2505",
    description="An agent that solves math problems and evaluates expressions.",
    name="Math Helper",
    instructions="You are a helpful math assistant. You can explain concepts, solve equations, and evaluate math expressions using the code interpreter.",
    tools=[{"type": "code_interpreter"}],
    completion_args={
        "temperature": 0.2,
        "top_p": 0.9
    }
)

Step 3: Interact with the Agent

Start a conversation asking the agent to solve a quadratic equation:

response = client.beta.conversations.start(
    agent_id=math_agent.id, inputs="Solve the quadratic equation 2x² + 3x - 2 = 0"
)
 
print(response)

Retrieve the final output and the executed code:

print(response.outputs[2].content)
print(response.outputs[1].info['code'])

Step 4: Plotting Results

Continue the dialogue to plot the quadratic function:

response = client.beta.conversations.append(
    conversation_id=response.conversation_id, inputs="Plot the function f(x) = 2x² + 3x - 2"
)
 
file_id = response.outputs[2].content[0].file_id
file_bytes = client.files.download(file_id=file_id).read()
with open(f"image_generated.png", "wb") as file:
    file.write(file_bytes)

Display the generated image:

from IPython.display import Image, display
image_path = "image_generated.png"
display(Image(filename=image_path))

This approach maintains conversation context for smooth and natural interactions, showcasing the flexibility and power of Mistral Agents API in creating advanced AI assistants.