Skip to content

Mistral Setup Guide

This guide walks you through setting up Mistral AI models with Janito.

Overview

Mistral AI provides a comprehensive suite of models including: - General-purpose models: Mistral Small, Medium, and Large - Code-focused models: Codestral for code generation and completion - Development-focused models: Devstral for agentic software development

Prerequisites

  1. Python 3.8+ installed
  2. Janito installed (pip install janito)
  3. Mistral API Key from La Plateforme

Quick Setup

1. Get Your API Key

  1. Visit Mistral AI La Plateforme
  2. Sign up or log in to your account
  3. Navigate to API Keys section
  4. Create a new API key
  5. Copy the key for use in Janito

2. Configure Janito

Set your Mistral API key using one of these methods:

# Linux/macOS
export MISTRAL_API_KEY="your-api-key-here"

# Windows (PowerShell)
$env:MISTRAL_API_KEY="your-api-key-here"

# Windows (Command Prompt)
set MISTRAL_API_KEY=your-api-key-here

Method 2: Janito CLI

janito set-api-key mistral your-api-key-here

Method 3: Interactive Setup

janito chat
# Then use the !model command to select mistral and follow prompts

Available Models

General Purpose Models

  • mistral-large-latest (default) - Most capable with 128k context
  • mistral-medium-latest - Balanced performance with 32k context
  • mistral-small-latest - Compact and efficient with 32k context

Code-Focused Models

  • codestral-latest - Specialized for code generation with 256k context
  • codestral-2405 - Previous version of code-focused model

Development-Focused Models

  • devstral-small-latest - Optimized for agentic tool use
  • devstral-medium-latest - Enhanced agentic capabilities

Usage Examples

Basic Chat

# Start chat with default model (mistral-large-latest)
janito chat --provider mistral

# Use specific model
janito chat --provider mistral --model codestral-latest

Single-Shot Mode

# Quick question
janito "Explain quantum computing" --provider mistral

# Code generation
janito "Write a Python function for fibonacci" --provider mistral --model codestral-latest

Configuration File

You can also configure Mistral in your Janito configuration:

# ~/.janito/config.yaml
providers:
  mistral:
    api_key: "your-api-key-here"
    model: "mistral-large-latest"

Model Selection Tips

  • General conversations: Use mistral-large-latest for best results
  • Code generation: Use codestral-latest for programming tasks
  • Development workflows: Use devstral-* models for agentic development
  • Cost optimization: Use mistral-small-latest for simpler tasks

Troubleshooting

Common Issues

  1. "Invalid API key" error
  2. Verify your API key is correct
  3. Check if the key has proper permissions

  4. Ensure no extra spaces in the key

  5. "Model not found" error

  6. Check available models with: janito list-models --provider mistral

  7. Ensure you're using the exact model name

  8. Rate limiting

  9. Mistral has rate limits based on your plan
  10. Consider upgrading your plan or implementing retry logic

Debug Mode

Enable debug logging to see API requests:

# Enable HTTP debug for Mistral
export MISTRAL_DEBUG_HTTP=1
janito chat --provider mistral --verbose-api

Testing Connection

Test your setup:

# Ping the provider
janito ping-providers

# List available models
janito list-models --provider mistral

# Test with a simple query
janito "Hello, Mistral!" --provider mistral

Advanced Configuration

Custom Base URL

If you need to use a different endpoint (e.g., for enterprise deployments):

janito set-config mistral.base_url "https://your-custom-endpoint.com/v1"

Model Parameters

You can set default parameters for Mistral models:

janito set-config mistral.temperature 0.5
janito set-config mistral.max_tokens 2000

Support

For additional help: - Check the Mistral AI documentation - Review Janito troubleshooting guides - Join our Discord community